Research ArticleA Novel Blind and Robust Video Watermarking Technique inFast Motion Frames Based on SVD and MR-SVD

Imen Nouioua 1 Nouredine Amardjia2 and Sarra Belilita1

1LCCNS Laboratory Department of Electronics Faculty of Technology Ferhat Abbas Setif University Algeria2LIS Laboratory Department of Electronics Faculty of Technology Ferhat Abbas Setif University Algeria

Correspondence should be addressed to Imen Nouioua nouioua imeneuniv-setifdz

Received 23 July 2018 Revised 6 October 2018 Accepted 14 October 2018 Published 13 November 2018

Academic Editor David Megias

Copyright copy 2018 Imen Nouioua et al This is an open access article distributed under the Creative Commons Attribution Licensewhich permits unrestricted use distribution and reproduction in any medium provided the original work is properly cited

In this work a novel and efficient digital video watermarking technique based on the Singular Value Decomposition performed inthe Multiresolution Singular Value Decomposition domain is proposed While most of the existing watermarking schemes embedthe watermark in all the video frames which is time-consuming and also affects the perceptibly of the video quality the proposedmethod chooses only the fast motion frames in each shot to host the watermark In doing so the number of frames to be processedis consequently reduced and a better quality of the watermarked video is also ensured since the human visual system cannot noticethe variations in fast moving regions The watermark information is embedded by Quantization IndexModulation which is a blindwatermarking algorithm The experimental results demonstrate that the proposed method can achieve a very good transparencywhile being robust against various kinds of attacks such as filtering noising compression and frame collusion Compared withseveral methods found in the literature the proposed method gives a better robustness

1 Introduction

In recent years the fabulous growth of the internet technol-ogy and the expansion of powerful computing devices havenot only boosted the multimedia electronic commerce up butalso incited artists to share and promote their work onlineThis obviously implied a massive presence on the web of dig-ital multimedia data such as audio image and video How-ever with the spread-out and ease of use of powerful multi-media dedicated processing tools these data can be down-loaded easilymodified illicitly appropriated and then largelyredistributed or commercialized on the Internet Protectingintellectual property rights of owners has then become amajor concern A solution to this problem is provided by digi-tal watermarking [1]

A secure digital watermarking technique comprises twoprocedures an embedding procedure and an extractionprocedure performed by the use of embedding and extractionkeys The embedding procedure consists of inserting in thehost multimedia content (usually called the cover) a water-mark which is a digital signature that holds copyright infor-mation exclusively limited to the owner Consequently by

means of the given secret keys the extraction procedure per-mits solely to the owner or to an authorized recipient of thedigital content to retrieve the watermark from the water-marked content [2]

In an efficient watermarking process two important pro-perties have to be taken into account [3] (1) imperceptibilityfor an invisible watermarking scheme there must be no dis-cernible difference between the original and the watermarkedcontents and (2) robustness the embeddedwatermark shouldbe able to survive to some extent intentional and uninten-tional content manipulations

Digital watermarking systems work either in the spatialdomain or in a transform domain A spatial domain tech-nique works directly on pixels the watermark is embeddedby usually modifying directly the pixels values such as leastsignificant bits (LSBs) [4] whereas a transform domain tech-nique embeds the watermark by adjusting the transformdomain coefficients Popular transforms that have been fre-quently used are the Discrete Cosine Transform (DCT) [5]the Discrete Wavelet Transform (DWT) [6] the DiscreteFourier Transform (DFT) [4] and the Singular ValueDecom-position (SVD) [7 8] Many combinations between these

HindawiSecurity and Communication NetworksVolume 2018 Article ID 6712065 17 pageshttpsdoiorg10115520186712065

2 Security and Communication Networks

transforms have also been investigated in the literature toaccomplish better results [9 10] Compared to spatial domaintechniques transform domain ones have been shown toachieve better robustness and imperceptibility [9] Further-more according to the watermark extracting process digitalwatermarking systems are categorized in three schemes [2]blind semi-blind and non-blind In a blind watermarkingscheme neither the original cover nor the embedded water-marks are required for detection but just the secret keys[2 7 11] In a semi-blind watermarking scheme only someinformation from the original cover and the secret keys areneeded [2 12] A non-blind watermarking scheme requiresthe original cover the original watermark and the secret keys[2 9] This makes the blind watermarking schemes the mostchallenging ones to develop

Initially digital watermarking has beenmainly studied forstill images but in recent few years a considerable numberof techniques dealing with video watermarking have beenconsidered However one must say that video watermarkingalgorithms are more difficult to develop than those operatingon images This is essentially due to the temporal dimensionwhich necessitates some specific requirements [13] (1)Therobustness of the watermark should deal not only with com-mon image processing attacks such as noise adding and JPEGcompression but also with video processing attacks such asMPEG compression and frame synchronization attacks (2)The imperceptibility in video watermarking is more difficultto achieve due to motion of objects in video sequences sothe temporal dimension should be taken into account inorder to avoid distortion between frames (3)The complexityof the watermarking scheme should be low because of thesignificant number of frames to be processed in a video signalGiven that a digital video sequence is considered basicallyas a collection of sequential images [14] many of the imagewatermarking techniques that are present in the literaturewere extended to video [6 9 15 16] as they embed thewatermark in all frames of the video sequences Thus thesealgorithms are robust to frame dropping and frame swappingbut in return they are time-consuming and also affect theperceptibly of the video quality To solve this problem offrame by frame embedding an answer to the following keyquestion should be found What are the preferred framesto host the watermark without degrading the visual qualityof the watermarked video while maintaining the robustnessreasonably unaffected The answer is to adaptively embedthe watermark in selected frames In this direction very fewvideowatermarking schemeswere considered Tabassum andIslam [17] proposed a digital video watermarking techniquebased on identical frame extraction In this method thehost video is initially divided into video shots Then fromeach video shot one video frame called identical frame isselected for watermark embedding In [18] Agilandeeswariand Ganesan developed an approach for video watermarkingusing SVD and DWT In their algorithm they extractedthe non-motion frames from the video using histogramdifference based scene change detection algorithm and thenthey embedded in them the same watermark However theproblem in these techniques is the small number of water-marked video frames So if those embedded frames are lostthe scheme becomes unreliable

Jiang Xuemei et al [19] developed an approach for videowatermarking based on shot segmentation and block classi-fication They selected the frames with the biggest luminancevalue in every shot to be the host frames The watermark sig-nal is cropped into small watermarks according to the num-ber of host frames in the host video These small watermarksare then respectively embedded into the different selectedhost frames Also Chetan et al [20] proposed a robust videowatermarking scheme based on scene changes which embeddifferent parts of a single watermark into different scenes of avideoThese frames are selected based on scene change detec-tion In these two last cited techniques if one watermarkedframe is lost the watermark cannot be extracted completely

In this work we propose a novel video watermarkingscheme in fast motion frames using Singular Value Decom-position in the Multiresolution Singular Value Decomposi-tion (MR-SVD) domain The main contribution of our workis as follows

(i) In order to avoid embedding the watermark in all theframes of the video sequences we first segment thevideo into temporally stationary signals using shotboundary detection Then from each shot we choosethe frames with big motion energy (fast motionframes) to embed thewatermarkThis is done becausethe human visual system (HVS) cannot notice the de-tails of fast moving regions [21] and thus the percep-tual invisibility of the watermark is guaranteed

(ii) Because of their relevant advantages we use a com-bination of the SVD and MR-SVD transforms SVDwith its attractive mathematical properties has beenbroadly applied in image compression and imagewatermarking and proved to be an efficient techniquein both domains [22]Most existing SVDbasedwater-marking techniques combine the SVD transformwiththe multiresolution 2D-DWT [9 10] as they wereshown to be reliable and provide high robustnessand better perceptual image quality However oneof the drawbacks of the DWT is its huge resourcesconsummation and high computation cost due to theconvolutions carried out in each of the filters Toovercome this issue Kakarala and Ogunbona [23]proposed the idea of the MR-SVD which performsmultiresolution decomposition similar to that of thedwt has perfect reconstruction and above all is amatrix based operation like the SVD Therefore ahybrid SVD MR-SVD watermarking technique isbased only on matrix operations which make itwell suited for real-time applications and simple forhardware implementation

(iii) Also we embed watermark information by Quantiza-tion Index Modulation (QIM) which has been shownto be host interference free and provably optimal interms of channel capacity under an additive whiteGaussian noise attack Furthermore the extractionprocedure in QIM is blind whichmakes it suitable forrobust watermarking [24]

(iv) Moreover to embed the watermark in a secure man-ner we encrypt the watermark using a logistic mapbased encryption [25]

Security and Communication Networks 3

This paper is organized in five sections The next one intro-duces the preliminaries of our scheme Section 3 gives the de-tails of the proposed video watermarking which includefour parts the fast motion frames extraction the watermarkpreprocessing and the watermark embedding and extractingprocesses The experimental results concerning the trans-parency and robustness against various attacks with compar-isons with other previous algorithms found in the literatureare presented in Section 4 Finally conclusions are given inthe last section

2 Preliminaries

21 Singular Value Decomposition In linear algebra SingularValue Decomposition (SVD) is a numerical technique thatdecomposes a matrix into three matrices with valuableproperties when applied in digital image processing [22] Ifa matrix A represents for example an image of size N times Nthen the SVD of A is given by119860 = 119880119878119881119879= (11990611 sdot sdot sdot 1199061119873 d

1199061198731 sdot sdot sdot 119906119873119873)(1199041 sdot sdot sdot 0 d

0 sdot sdot sdot 119904119873)(V11 sdot sdot sdot V1119873 d

V1198731 sdot sdot sdot V119873119873)

(1)

where U and V are orthogonal matrices representing respec-tively the horizontal and vertical details (edges) of the imageand S is a diagonal matrix where the diagonal elementsSi (1 le i le N) with S1 ge S2 ge sdot sdot sdot ge SN are the singularvalues (SVs) of A

Two main properties related to the SVs make the SVDappropriate for watermarking when the matrix S is utilized[8 22]

(i) The energy content (luminance) of the image A islocated in the SVs

(ii) The SVs have very good stability ie a small pertur-bation added (a watermark for example) to the imagedoes not change significantly the SVs

22Multiresolution Singular ValueDecomposition (MR-SVD)As stated in the Introduction the MR-SVD initially intro-duced in [23] is a matrix based operation

221 1D Multiresolution Singular Value Decomposition LetX = [x(1) x(N)] represent a finite extent 1D signal andassume that N is divisible by 2L for some L ge 1 Let the datamatrix at the first level denoted by X1 be constructed withits top and bottom rows containing respectively the odd-numbered and even-numbered samples of X

1198831 = (119909 (1) 119909 (3) 119909 (119873 minus 1)119909 (2) 119909 (4) 119909 (119873) ) (2)

Thecorrespondingcentredmatrix is1198831 = 1198831(119868119873minus(1119873)119890119873119890119879119873where 119868119873 is the identity and eN is the vector containing allones

Let U1 be the eigenvector matrix bringing the scattermatrix 1198791 = 11988311198831198791 into diagonal form

UT1T1U1 = S21 (3)

where S21 = diags1(1)2 s1(2)2 contains the squares of thetwo singular values with s1(1) ge s1(2)

Now let X1 = UT1X1

The top row of X1 denoted by 01 = X1(1 ) correspondsto the largest eigenvalue and represents the approximationcomponent The bottom row of X1 designated by 1205931 =X1(2 ) corresponds to the smallest eigenvalue and containsthe detail component The successive levels of decompositionrepeat the procedure described above by placing the approx-imation component 01 in place of X Hence the MR-SVD canbe written as 119883 997888rarr 0119871 120593119897119871119897=1 119880119897119871119897=1 (4)

where L is the desired level of decomposition

222 2D Multiresolution Singular Value Decomposition Webriefly describe here the 2DMR-SVDThe first-level decom-position of the image proceeds as followsTheM timesN image Xis divided into nonoverlapping 2 times 2 blocks and each block isarranged into a 4 times 1 vector by stacking columns to form thedata matrix X1The eigendecomposition of the 4 times 4 scattermatrix is 1198791 = 11988311198831198791 = 1198801119878211198801198791 (5)

Let 1198831 = 11988011987911198831 (6)

The top row of the resulting matrix X1(1 ) is rearranged toform anM2timesN2matrix which is considered as the smooth(approximation) components of the image The remainingrows X1(2 ) X1(3 ) and X1(4 ) contain the detail com-ponents which are denoted by 1205931l 1205932l 1205933l respectively Thecomplete transform can be represented as follows119883 997888rarr 0119871 1205931119897 1205932119897 1205933119897 119871119897=1 119880119897119871119897=1 (7)

The original image X can be reconstructed from the righthand side since the steps are reversible As an examplethe one-level MR-SVD decomposition of the video frameldquoForemanrdquo is depicted in Figure 1

3 Proposed Method

Our proposed algorithmencompasses four consecutive partsfast motion frames where to embed the watermark water-mark preprocessing watermark embedding process andwatermark extraction process

31 Fast Motion Frames Extraction for Embedding (FMFE)In order to increase the quality of the watermarked video

4 Security and Communication Networks

Figure 1 Original video frame Foreman image and its 1-level MR-SVD

the proposed system exploits the characteristics of the humanvisual system (HVS) to select frames in which the watermarkis embedded effectively Because the HVS is less sensitive toerrors in regions with great motion we select the frames thathave big motion energy to be the host frames [21] In ourscheme to extract the fast motion frames the cover video Vis first converted into individual frames then shot bound-aries (SBD) are detected using the algorithm proposed in[26] to obtain temporally stationary signals because frameswithin the same shot have a strong correlation Afterwardswe measure the average motion energy of each frame usingthe mean magnitudes of motion vectors (MMMV)

Let S = F1 F2 Fk be a shot of length k where Fk (k =1 2 k) represents the kth frame in the shot The MMMVof the frame can be calculated as follows119872119872119872119881(119865119895) = 1119873 sum

1le119894le1198731le119895le119896

1198721198812119909 (119894 119895) + 1198721198812119910 (119894 119895) (8)

where N is the number of macro blocks in a frame and MVxand MVy represent the components of the motion vector(MV) in respectively the X axis direction and the Y axisdirection

119879ℎ119903119890119904ℎ = 120572119896 119896sum119895=1 119872119872119872119881(119865119895)) (9)

The above threshold is used as a decision rule to distinguishbetween fast and slow motion frames Furthermore we use aconstant 120572 to adjust the decision rule

119865119895 119894119904 119891119886119904119905119890 119898119900119905119894119900119899 119891119903119886119898119890 119894119891 119872119872119872119881(119865119895) gt 119905ℎ119903119890119904ℎ119865119895 119894119904 119899119900 119898119900119905119894119900119899 119891119903119886119898119890 119894119891 119872119872119872119881(119865119895) le 119905ℎ119903119890119904ℎ (10)

Figure 3 shows the average motion energy of some frames inForeman Football and Akiyo video shots In the case of theForman video we see that the fast motion frames take placebetween the frames 150 and 228

32 Watermark Preprocessing In order to improve the secu-rity of the proposed algorithm the binary watermark shouldbe first preprocessed before embedded Here the watermark

is scrambled by using the chaotic logistic map which is deter-mined by the equation [25 27]119883119899+1 = 120582119883119899 (1 minus 119883119899) (11)

where 375 lt 120582 lt 4 is the system parameter The initial valueX0 L [0 1] is adopted as a key

Then the binary image logo or signature W(n) is scram-bled by X(n) with the following rule119882119875 (119899) = 119882 (119899) oplus 119883 (119899) 0 le 119899 le 119873 (12)

with N being the total number of bits in the watermark and oplusbeing the binary Exclusive Or (XOR) operation

33 Watermark Embedding Process In this section we de-scribe the proposed video watermarking scheme Figure 2shows the block diagram of the proposed video watermarkembedding procedure which is described as follows

(1) The fastmotion frames are extracted from the originalcolor video Only these frames are watermarked Thismakes the watermarked video quality good becausethe watermark is not embedded in all the frames as itis done in other watermarking schemes [6 9 28]

(2) Every fast motion frame is converted from the RGBto the YCbCr color space

(3) Every luminance frame is transformed with the 1-Level MR-SVD decomposition to get approximationand detail components 0 120593 1 1205932 1205933

(4) The approximation componentΦ is decomposed intoblocks of size [u times u]

(5) TheSVD is applied to each block of the approximationcomponent 0 (ie the low frequency subband) whichcontains the major video frame energy0119896119899 = 119880119896119899 times 119878119896119899 times 119881119879119896119899 (13)

where k is the fast motion frame index and n is thelocation of the block We use the SVD due to its sui-table properties discussed earlier

Security and Communication Networks 5

Watermark embedding process

Original video SVDOne level

MR-SVDutimesu

blocksRGB to YCbCr

FMFESBD

Watermark Chaotic Logistic Map Encryption

Watermarked video

YCbCr to RGB

Inverse MR-SVD

Compose utimesu blocks

Inverse SVD

３ＥＨ(11) =[ ３ＥＨ(11)

１]]１ +

3

4１ ７０(Ｈ) = 1

[ ３ＥＨ(11)１]１ +

1

4１ ７０(Ｈ) = 0

(a)

Watermark extraction process

Watermarked video

Recovered watermark

SBD FMFE SVDutimesu blocks

One level MR-SVD

RGB to YCbCr

Chaotic Logistic Map

decryption

７０

1 ３(11) minus １lfloor lfloor３

(11)

１ge

１

2

0 ３(11) minus １lfloor lfloor３

(11)

１lt

１

2

(b)

Figure 2 Proposed video watermarking technique (a) embedding process and (b) recover process

(6) The watermark is encrypted using chaotic logisticmap

(7) In order to guarantee robustness to our watermarkingscheme the watermarking system inserts the water-mark bits in the largest singular value using the QIMmethod as

119878119896119899 (1 1) = [119878119896119899 (1 1)119876 ]119876 + 34119876 119882119875 (119899) = 1[119878119896119899 (1 1)119876 ]119876 + 14119876 119882119875 (119899) = 0 (14)

where Q is the quantization step and [∙] stands for therounding operation

(8) Inverse SVD transformation is conducted to obtainthe watermarked block01015840119896119899 = 119880119896119899 times 1198781015840119896119899times119881119879119896119899 (15)

(9) To build the watermarked luminance inverse MR-SVD is applied to the modified approximation com-ponent 01015840

(10) Reconstruction of the watermarked video frame isdone using the watermarked luminance part and theoriginal frame chrominance parts Cb and Cr byconversion from the YCbCr into the RGB color space

34 Watermark Extraction Process The watermark extrac-tion process is shown in Figure 2 and is described as follows

(1) The watermarked fast motion frames are extractedfrom the watermarked color video

(2) Every watermarked fast motion frame is convertedfrom the RGB to the YCbCr

(3) Every watermarked luminance frame is transformedinto 1-LevelMR-SVD decomposition to get the water-marked approximation 01015840

(4) The watermarked approximation component is de-composed into blocks of size [u times u]

6 Security and Communication Networks

0

1

2

3

4

5

6

7The average motion energy in the second shot of video ForemanaviThe average motion energy in the first shot of video Foremanavi

220 240 260 280200Frame Number

20 40 60 80 100 120 140 160 1800Frame Number

0

1

2

3

4

5

6

7

MM

MV

MMMVThresh

MMMVThresh

(a)

4

5

The average motion energy in the second shot of video Fooballavi The average motion energy in the first shot of video Fooballavi

0

05

1

15

2

25

3

35

4

45

5

MM

MV

36

38

42

44

46

48

52

54

56

MM

MV

275 280 285 290 295270Frame Number

50 100 150 200 2500Frame Number

MMMVThresh

MMMVThresh

(b)

Average motion energy in Akiyoavi

0005

01015

02025

03035

MM

MV

50 100 150 200 250 3000Frame Number

MMMVThresh

(c)

Figure 3 Average motion energy in (a) Foreman (b) Football and (c) Akiyo videos

Security and Communication Networks 7

(5) TheSVD is applied to each block of the approximationcomponent01015840119896119899 = 119880119896119899 times 1198781015840119896119899times119881119879119896119899 (16)

(6) The embedded watermark is extracted by the follow-ing rule

119875 = 1 1198781015840 (1 1) minus 119876lfloor1198781015840 (1 1)119876 rfloor ge 11987620 1198781015840 (1 1) minus 119876lfloor1198781015840 (1 1)119876 rfloor lt 1198762 (17)

where lfloor∙rfloor is the floor function(7) Decryption with the same chaotic sequence is per-

formed to get the hidden binary watermark W

(8) Since a video clip contains several fast motion framesin which the same watermark is embedded we calcu-late the final recovered watermark by averaging thewatermarks extracted from these different frames

4 Results and Discussion

Theproposed algorithm is implemented using MATLABWeused three CIF (352times288) standard sequences shown in Fig-ure 4 (Foreman 300 frames Akiyo 300 frames and Football260 frames) The test videos are in RGB uncompressed aviformat with a frame rate of 30 fps [29] The watermark is abinary image Its resolution is 9 times 11 The block size in theproposed algorithm can be set to any desired value Howevera small block size leads to block effects problem whereas alarge block size reduces the total number of the watermarkbits We carried out various experiments and found that theblock size of 16x16 allows the embedding of an acceptablenumber of watermark bits without causing noticeable blockeffect

As shown inFigure 5 small values of the quantization stepyield good transparency at the expense of poor robustnessand vice versa From this figure it can be observed that thevalue of the quantization step Q = 110 gives a good com-promise between robustness and imperceptibility

41 Imperceptibility Tests The imperceptibility of the water-mark is estimated by measuring the PSNR (Peak Signal toNoise Ratio) and Mean Structure Similarity Index Measure(MSSIM) which are calculated using the luminance space Yof the original and watermarked frames [30] High PSNRvalues of the watermarked video frames indicate betterimperceptibility It is worth noting that the SSIM provides aperceptual distortion in range of [0 1] When both framesare numerically the same this value is equal to 1 The PSNRis calculated as follows

119875119878119873119877 = 10 log10 ( 2562119872119878119864) (18)

Table 1 The PSNR and MSSIM of the watermarked videos

Video Foreman Akiyo FootballAverage PSNR 4020 4045 4026Average MSSIM 09981 09988 09977

where the Mean Square Error (MSE) between the host lumi-nance Y and the watermarked luminance Y1015840 is defined as

119872119878119864 = 1119872 times119873119872minus1sum119894=0 119873minus1sum119895=0 1003816100381610038161003816119884 (119894 119895) minus 1198841015840 (119894 119895)100381610038161003816100381610038162 (19)

with M and N respectively being the height and width of thevideo frame

The MSSIM is defined as follows

119872119878119878119868119872(119884 1198841015840) = 1119872 119872sum119895=1 119878119878119868119872(119884119869 1198841015840119895) (20)

where YJ and Y1015840j are the image contents at the j local windowand M is the number of local windows of the image

119878119878119868119872(119884 1198841015840) = [119897 (119884 1198841015840)]120572 sdot [119888 (119884 1198841015840)]120573 sdot 119904 [119884 1198841015840)]120574 (21)

where l c and s are the luminance comparison the contrastcomparison and the structure comparison functions respec-tively With 120572 120573 and 120574 are parameters used to adjust therelative importance of the three components

As shown inTable 1 the average PSNRvalues of thewater-marked videos are higher than 40 dB and the correspondingMSSIMvalues are very close to 1This indicates the invisibilityof the watermark which means that the watermarked videosappear visually identical to the original ones as shown inFigure 6

42 Robustness Tests The robustness for any watermarkingsystem is a very important requirement To verify it weapply to the watermarked video various types of attacks andwe use the Normalized Coefficient (NC) and the Bit ErrorRate (BER) to compare the similarities between the originalwatermark W and the extracted watermark W The NC andBER are respectively calculated as

119873119862(119882 ) = sumsum 119882(119894 119895) (119894 119895)1003816100381610038161003816119882 (119894 119895)10038161003816100381610038162 (22)

119861119864119877 = 1119875 119901sum119895=1

10038161003816100381610038161003816 (119895) minus 119882 (119895)10038161003816100381610038161003816 (23)

where W W and P are respectively the original watermarkthe extracted watermark and the size of the watermark

The correlation between W and W is very high when NCis close to 1

Generally in video watermarking attacks are divided intothree categories image processing attacks frame synchro-nization attacks and video compression attacks

8 Security and Communication Networks

(a) (b)

(c)

Figure 4 Video sequences used for testing (a) Foreman (b) Akiyo and (c) Football

1

09965

0997

09975

0998

09985

0999

09995

MSS

IM

60 80 100 120 14040Q-step

(a)

60 80 100 120 14040Q-step

055

06

065

07

075

08

085

09

095

1

NC

(b)

Figure 5 (a) MSSIM of watermarked video for different quantization step Q and (b) NC of extracted watermark with JPEG compressionattack for different quantization step Q

Image Processing Attacks Considering a video as a sequenceof images the attacks applied to images can then be applied tothe video sequences The common image processing attacksare as follows

(i) Adding a noise three kinds of noises are added tothe watermarked video Gaussian noise salt amp peppernoise and speckle noise with density of 1 It canbe seen from Table 2 that the watermark is always

Security and Communication Networks 9

(a) (b)

(c)

Figure 6 Original and watermarked frames (a) frame 228 of Foreman (b) frame 16 of Akiyo and (c) frame 128 of Football

detectablewithNCandBER values respectively closeto 1 and 0 especially for Foreman and Football videosequences

(ii) Filtering 3x3median filter and 5x5 Gaussian filter areapplied separately to the watermarked videos and wecan see from Table 2 that the proposed method isrobust against median and Gaussian filtering

(iii) JPEG compression the watermarked video is com-pressed with different quality factors ranging from 10to 100 Figure 7 shows the results for the JPEG com-pression for instance if the watermarked video iscompressed with a quality factor of 40 the obtainedNC is greater than 95 and the BER value is lowerthan 02 This confirms the robustness of the pro-posed scheme to the JPEG compression attack

Frame Synchronization Attacks Because contents in the con-secutive frames of a video are almost identical it makesthe video sequences susceptible to temporal synchronizationattacks such as frame averaging frame dropping and frameswapping

(i) Frame dropping in frame dropping selected water-marked frames are replaced by their correspondingoriginal frames Table 3 shows the average NC andBER values given at different frame dropping ratesOur scheme achieves strong robustness against framedropping even for the case of high rates (ie 80)

(ii) Frame averaging in frame averaging we replaceselected watermarked frames by the average of theirprevious current and next frames The watermarkedvideo is averaged for various averaging rates and thenwe tried to extract the watermark Table 4 shows thatthe watermark can be recovered at frame averagingrates up to 50

(iii) Frame swapping the results presented in Table 5prove the robustness of our watermarking schemeagainst frame swapping because when all water-marked frames are swapped the NC is 1 and the BERis 0

Video Compression Attacks Video compression is a funda-mental attack in video watermarking that should be verifiedas video sequences are stored and transmitted in compressedformat Here we use a tool for video processing namedVirtualDub to compress the videos sequences with twodifferent Lossy compressions [31] H264 coding with a bit rateof 512kbps andMPEG4 coding with bit rates of 1500kbps and1000kbpsTheNC and BER are depicted in Table 6 and we seethat the algorithm can resist to video compression attacks

Results with the StirMark Benchmark We also evaluated theproposed method with StirMark 31 which is a well-knownevaluation tool for watermarking robustness of watermarkedvideo frames under image processing attacks [32 33] Table 7shows the evaluation of the proposed method with theStirMark benchmark under JPEG compression median filterwith JPEG compression Gaussian filter with JPEG compres-sion sharpening with JPEG compression and removal ofrow and clone with JPEG compression As can be observedthe proposed watermarking technique can resist to all thesenamed attacks

43 Comparison with Some Previously Reported AlgorithmsIn order to evaluate the performance of our algorithm wecompared the results of the proposed video watermarkingscheme to the results of related video watermarking schemesgiven in [6 16 18 20] which we introduced and discussedin the Introduction Figure 8 displays these results of com-parison under frame dropping frame averaging and JPEG

10 Security and Communication Networks

Table2Ex

tractedwatermark(w

ithNCandBE

Rvalues)u

nder

imagep

rocessingattack

Attacks

Foreman

Akiyo

Football

Noattack

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

Gaussianno

ise

NC=090

08BE

R=004

04NC=066

74BE

R=01414

NC=1

BER=0

SaltampPepp

erno

ise

NC=1

BER=0

NC=09238

BER=0303

NC=1

BER=0

Specklen

oise

NC=09479

BER=00202

NC=07870

BER=009

09NC=1

BER=0

Medianfilter

NC=09732

BER=0101

NC=1

BER=0

NC=1

BER=0

Gaussianlowpassfilter

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

Security and Communication Networks 11

Table3Ex

tractedwatermark(w

ithNCandBE

R)un

derframed

ropp

ingattack

Fram

edropp

ingrate

Foreman

Akiyo

Football

50

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

80

NC=08485

BER=006

06NC=07822

BER=0101

NC=07156

BER=01313

12 Security and Communication Networks

Table4Ex

tractedwatermark(w

ithNCandBE

R)un

derframea

veraging

attack

Fram

eaveraging

rate

Foreman

Akiyo

Football

25

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

50

NC=06819

BER=0111

1NC=0800

BER=009

09NC=1

BER=0

Security and Communication Networks 13

Table5Ex

tractedwatermark(w

ithNCandBE

R)un

derframes

wapping

attack

Fram

eSwapping

rate

Foreman

Akiyo

Football

50

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

100

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

14 Security and Communication Networks

Table6Ex

tractedwatermark(w

ithNCandBE

R)un

derv

ideo

compressio

nattack

Foreman

Akiyo

Football

H264compressio

n(512kbps)

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

MPE

G4compressio

n(1500kbp

s)NC=1

BER=0

NC=08708

BER=00505

NC=1

BER=0

MPE

G4compressio

n(100

0kbp

s)NC=1

BER=0

NC=0848

BER=006

06NC=09238

BER=00303

Security and Communication Networks 15

100

90

80

70

60

50

4010 20 30 40 50 60 70 80 90 100

NC

()

JPEG quality factor

10 20 30 40 50 60 70 80 90 100

JPEG quality factor

ForemanAkiyoFootball

ForemanAkiyoFootball

30

25

20

15

10

5

0

BER

()

Figure 7 NC and BER values of the extracted watermarks under the JPEG compression attack

Table 7 Extracted watermark (with NC and BER) under image processing attacks with StirMark benchmark

Attacks NC BERJPEG 10 05677 01515JPEG 30 09479 00202JPEG 60 08485 00606JPEG 90 10000 000003x3 median filter+ JPEG 09732 00101Gaussian filter +JPEG 07681 01010Sharpening +JPEG 05127 02626removed one row one colon +JPEG 05523 02020

compression attacks It is clear that the proposed schemeachieves a better robustness

44 Complexity of the Proposed Algorithm The proposedwatermarking schema has a low complexity because of thefollowing

(i) The algorithm is done in the MR-SVD domain whichhas an overall complexity O(n) where n indicates thesignal length [23]

(ii) To extract fast motion frames we use a simple thresh-old decision method based on motion energy

(iii) The embedding and extraction are not applied to allthe frames of the video but just to fast motion frames

In fact the execution time of the proposed scheme required inwatermark embedding and extraction processes the followingresults watermark embedding 036 s per frame watermarkextraction 019 s per frame The simulation is conducted on

Intel(R) Core i5 180 GHz processor with 4 GB RAMusingMATLAB verR2013b

5 Conclusion

In this paper we proposed a novel blind video watermarkingscheme in fast motion frames using the SVD the MR-SVD the QIM and the Logistic Map encryption In theproposed technique we solved the problem of embedding thewatermark in all frames by choosing only the frames that havebig motion energy to be the host frames suitable to the HSVUsing this approach we have limited the number of frames tobe processed and also assured a higher imperceptibility of thewatermark In addition the combination of the characteris-tics of the SVD MR-SVD QIM and Logistic Map Encryp-tion makes our scheme secure and robust to a variety ofattacks The experimental results confirm that the proposedwatermarking scheme has good imperceptibility with PSNRgreater than 40 dBTheproposed scheme is robust not only toimage processing attacks but also to frame synchronization

16 Security and Communication Networks

100

95

90

85

80

75

NC

()

NC

()

100

95

90

85

80

75

NC

()

10 15 20 25 30 35 40

frame averaging rate

proposed method[20][16][6]

proposed method

proposed method

[20]

[20]

[16][6]

[6]

[18]

10 20 30 40 50 60 70

frame dropping rate

70

65

100

95

90

85

80

75

70

65

60

JPEG Quality factor20 30 40 50 60 70 80 90

Figure 8 Comparison of experimental results between the proposed technique and other schemes under frame averaging attack framedropping attack and JPEG compression attack

and video compression attacks The comparison results withother algorithms related to video watermarking indicate thesuperiority of our scheme

Data Availability

The data used to support the findings of this study are avail-able from the corresponding author upon request

Conflicts of Interest

The authors declare no conflicts of interest

References

[1] J C IngemarM LMiller A B Jeffrey J Fridrich andT KalkerDigital Watermarking and Steganography Morgan KaufmannPublishers Inc 2008

[2] S Katzenbeisser and F Petitcolas ldquoInformation Hiding Tech-niques for Steganography and Digital Watermarkingrdquo ArtechHouse 2000

[3] H Tao L Chongmin J M Zain and A N Abdalla ldquoRobustimagewatermarking theories and techniques A reviewrdquo Journalof Applied Research and Technology vol 12 no 1 pp 122ndash1382014

Security and Communication Networks 17

[4] M Kutter F Jordan and F Bossen ldquoDigital watermarking ofcolor images using amplitude modulationrdquo Journal of ElectronicImaging vol 7 no 2 pp 326ndash332 1998

[5] A Karmakar A Phadikar B S Phadikar and G K Maity ldquoAblind video watermarking scheme resistant to rotation and col-lusion attacksrdquo Journal of King Saud University - Computer andInformation Sciences vol 28 no 2 pp 199ndash210 2016

[6] T R Singh K M Singh and S Roy ldquoVideo watermarkingscheme based on visual cryptography and scene change detec-tionrdquo AEU - International Journal of Electronics and Communi-cations vol 67 no 8 pp 645ndash651 2013

[7] W Kong B Yang D Wu and X Niu ldquoSVD Based Blind VideoWatermarking Algorithmrdquo in Proceedings of the First Interna-tional Conference on Innovative Computing Information andControl - Volume I (ICICICrsquo06) pp 265ndash268 Beijing ChinaSeptember 2006

[8] R Liu and T Tan ldquoAn SVD-based watermarking scheme forprotecting rightful ownershiprdquo IEEE Transactions on Multime-dia vol 4 no 1 pp 121ndash128 2002

[9] O S Faragallah ldquoEfficient video watermarking based on sin-gular value decomposition in the discrete wavelet transformdomainrdquo AEU - International Journal of Electronics and Com-munications vol 67 no 3 pp 189ndash196 2013

[10] K Niu X Yang and L Xiang ldquoHybrid quasi-3D DWTDCTand SVD video watermarkingrdquo in Proceedings of the 2010 IEEEInternational Conference on Software Engineering and ServiceSciences ICSESS 2010 pp 588ndash591 China July 2010

[11] N I Yassin NM Salem andM I El Adawy ldquoQIM blind videowatermarking scheme based on Wavelet transform and prin-cipal component analysisrdquo Alexandria Engineering Journal vol53 no 4 pp 833ndash842 2014

[12] D K Thind and S Jindal ldquoA semi blind DWT-SVD videowatermarkingrdquo inProceedings of the International Conference onInformation and Communication Technologies ICICT 2014 pp1661ndash1667 India December 2014

[13] N Leelavathy E V Prasad and S Kumar ldquoVideo Watermark-ing Techniques A Reviewrdquo International Journal of ComputerApplications vol 104 no 7 pp 24ndash30 2014

[14] F Hartung and M Kutter ldquoMultimedia watermarking techni-quesrdquo Proceedings of the IEEE vol 87 no 7 pp 1079ndash1107 1999

[15] P Rasti S Samiei M Agoyi S Escalera and G AnbarjafarildquoRobust non-blind color video watermarking using QR decom-position and entropy analysisrdquo Journal of Visual Communica-tion and Image Representation vol 38 pp 838ndash847 2016

[16] S M Youssef A A ElFarag and N M Ghatwary ldquoAdaptivevideo watermarking integrating a fuzzy wavelet-based humanvisual system perceptual modelrdquo Multimedia Tools and Appli-cations vol 73 no 3 pp 1545ndash1573 2014

[17] T Tabassum and S M Islam ldquoA digital video watermarkingtechnique based on identical frame extraction in 3-LevelDWTrdquoin Proceedings of the 2012 15th International Conference onComputer and Information Technology (ICCIT) pp 101ndash106Chittagong Bangladesh December 2012

[18] L Agilandeeswari and K Ganesan ldquoA robust color video water-marking scheme based on hybrid embedding techniquesrdquoMul-timedia Tools and Applications vol 75 no 14 pp 8745ndash87802016

[19] X Jiang Q Liu and Q Wu ldquoA new video watermarking algo-rithm based on shot segmentation and block classificationrdquoMultimedia Tools and Applications vol 62 no 3 pp 545ndash5602013

[20] K Chetan and K Raghavendra ldquoDWT Based Blind DigitalVideo Watermarking Scheme for Video Authenticationrdquo Inter-national Journal of Computer Applications vol 4 no 10 pp 19ndash26 2010

[21] Y JiaW Lin and A A Kassim ldquoEstimating just-noticeable dis-tortion for videordquo IEEE Transactions on Circuits and Systems forVideo Technology vol 16 no 7 pp 820ndash829 2006

[22] H C Andrews and C L Patterson ldquoSingular value decomposi-tions and digital image processingrdquo IEEE Transactions on SignalProcessing vol 24 no 1 pp 26ndash53 1976

[23] R Kakarala and P O Ogunbona ldquoSignal analysis using a multi-resolution form of the singular value decompositionrdquo IEEETransactions on Image Processing vol 10 no 5 pp 723ndash7352001

[24] B Chen and G WWornell ldquoQuantization index modulation aclass of provably goodmethods for digital watermarking and in-formation embeddingrdquo IEEE Transactions on Information The-ory vol 47 no 4 pp 1423ndash1443 2001

[25] M Yaghoobi ldquoA New Approach for Image Encryption UsingChaotic Logistic Maprdquo in Proceedings of the 2008 InternationalConference on Advanced Computer Theory and Engineering(ICACTE) pp 585ndash590 Phuket Thailand December 2008

[26] R Dugad K Ratakonda and N Ahuja ldquoRobust video shotchange detectionrdquo in Proceedings of the 1998 IEEE SecondWork-shop on Multimedia Signal Processing pp 376ndash381 RedondoBeach CA USA 1998

[27] A Anees I Hussain A Algarni and M Aslam ldquoA RobustWatermarking Scheme for Online Multimedia Copyright Pro-tection Using New Chaotic Maprdquo Security and CommunicationNetworks vol 2018 2018

[28] R Thanki V Dwivedi and K Borisagar ldquoA hybrid watermark-ing scheme with CS theory for security of multimedia datardquoJournal of King Saud University - Computer and InformationSciences 2017

[29] httpmediaxiphorgvideoderf Accessed on 06 July 2018[30] B G Haskell and A N Netravali Digital Pictures Representa-

tion Compression and Standards Perseus Publishing 1997[31] httpwwwvirtualduborg Accessed on 06 July 2018[32] F A Petitcolas R J Anderson and M G Kuhn ldquoAttacks on

CopyrightMarking Systemsrdquo in Information Hiding vol 1525 ofLecture Notes in Computer Science pp 218ndash238 Springer BerlinHeidelberg Berlin Heidelberg 1998

[33] httpwwwpetitcolasnetwatermarkingstirmark Accessed on19 September 2018

Security and Communication Networks 3

This paper is organized in five sections The next one intro-duces the preliminaries of our scheme Section 3 gives the de-tails of the proposed video watermarking which includefour parts the fast motion frames extraction the watermarkpreprocessing and the watermark embedding and extractingprocesses The experimental results concerning the trans-parency and robustness against various attacks with compar-isons with other previous algorithms found in the literatureare presented in Section 4 Finally conclusions are given inthe last section

2 Preliminaries

21 Singular Value Decomposition In linear algebra SingularValue Decomposition (SVD) is a numerical technique thatdecomposes a matrix into three matrices with valuableproperties when applied in digital image processing [22] Ifa matrix A represents for example an image of size N times Nthen the SVD of A is given by119860 = 119880119878119881119879= (11990611 sdot sdot sdot 1199061119873 d

1199061198731 sdot sdot sdot 119906119873119873)(1199041 sdot sdot sdot 0 d

0 sdot sdot sdot 119904119873)(V11 sdot sdot sdot V1119873 d

V1198731 sdot sdot sdot V119873119873)

(1)

where U and V are orthogonal matrices representing respec-tively the horizontal and vertical details (edges) of the imageand S is a diagonal matrix where the diagonal elementsSi (1 le i le N) with S1 ge S2 ge sdot sdot sdot ge SN are the singularvalues (SVs) of A

Two main properties related to the SVs make the SVDappropriate for watermarking when the matrix S is utilized[8 22]

(i) The energy content (luminance) of the image A islocated in the SVs

(ii) The SVs have very good stability ie a small pertur-bation added (a watermark for example) to the imagedoes not change significantly the SVs

22Multiresolution Singular ValueDecomposition (MR-SVD)As stated in the Introduction the MR-SVD initially intro-duced in [23] is a matrix based operation

221 1D Multiresolution Singular Value Decomposition LetX = [x(1) x(N)] represent a finite extent 1D signal andassume that N is divisible by 2L for some L ge 1 Let the datamatrix at the first level denoted by X1 be constructed withits top and bottom rows containing respectively the odd-numbered and even-numbered samples of X

1198831 = (119909 (1) 119909 (3) 119909 (119873 minus 1)119909 (2) 119909 (4) 119909 (119873) ) (2)

Thecorrespondingcentredmatrix is1198831 = 1198831(119868119873minus(1119873)119890119873119890119879119873where 119868119873 is the identity and eN is the vector containing allones

Let U1 be the eigenvector matrix bringing the scattermatrix 1198791 = 11988311198831198791 into diagonal form

UT1T1U1 = S21 (3)

where S21 = diags1(1)2 s1(2)2 contains the squares of thetwo singular values with s1(1) ge s1(2)

Now let X1 = UT1X1

The top row of X1 denoted by 01 = X1(1 ) correspondsto the largest eigenvalue and represents the approximationcomponent The bottom row of X1 designated by 1205931 =X1(2 ) corresponds to the smallest eigenvalue and containsthe detail component The successive levels of decompositionrepeat the procedure described above by placing the approx-imation component 01 in place of X Hence the MR-SVD canbe written as 119883 997888rarr 0119871 120593119897119871119897=1 119880119897119871119897=1 (4)

where L is the desired level of decomposition

222 2D Multiresolution Singular Value Decomposition Webriefly describe here the 2DMR-SVDThe first-level decom-position of the image proceeds as followsTheM timesN image Xis divided into nonoverlapping 2 times 2 blocks and each block isarranged into a 4 times 1 vector by stacking columns to form thedata matrix X1The eigendecomposition of the 4 times 4 scattermatrix is 1198791 = 11988311198831198791 = 1198801119878211198801198791 (5)

Let 1198831 = 11988011987911198831 (6)

The top row of the resulting matrix X1(1 ) is rearranged toform anM2timesN2matrix which is considered as the smooth(approximation) components of the image The remainingrows X1(2 ) X1(3 ) and X1(4 ) contain the detail com-ponents which are denoted by 1205931l 1205932l 1205933l respectively Thecomplete transform can be represented as follows119883 997888rarr 0119871 1205931119897 1205932119897 1205933119897 119871119897=1 119880119897119871119897=1 (7)

The original image X can be reconstructed from the righthand side since the steps are reversible As an examplethe one-level MR-SVD decomposition of the video frameldquoForemanrdquo is depicted in Figure 1

3 Proposed Method

Our proposed algorithmencompasses four consecutive partsfast motion frames where to embed the watermark water-mark preprocessing watermark embedding process andwatermark extraction process

31 Fast Motion Frames Extraction for Embedding (FMFE)In order to increase the quality of the watermarked video

4 Security and Communication Networks

Figure 1 Original video frame Foreman image and its 1-level MR-SVD

the proposed system exploits the characteristics of the humanvisual system (HVS) to select frames in which the watermarkis embedded effectively Because the HVS is less sensitive toerrors in regions with great motion we select the frames thathave big motion energy to be the host frames [21] In ourscheme to extract the fast motion frames the cover video Vis first converted into individual frames then shot bound-aries (SBD) are detected using the algorithm proposed in[26] to obtain temporally stationary signals because frameswithin the same shot have a strong correlation Afterwardswe measure the average motion energy of each frame usingthe mean magnitudes of motion vectors (MMMV)

Let S = F1 F2 Fk be a shot of length k where Fk (k =1 2 k) represents the kth frame in the shot The MMMVof the frame can be calculated as follows119872119872119872119881(119865119895) = 1119873 sum

1le119894le1198731le119895le119896

1198721198812119909 (119894 119895) + 1198721198812119910 (119894 119895) (8)

where N is the number of macro blocks in a frame and MVxand MVy represent the components of the motion vector(MV) in respectively the X axis direction and the Y axisdirection

119879ℎ119903119890119904ℎ = 120572119896 119896sum119895=1 119872119872119872119881(119865119895)) (9)

The above threshold is used as a decision rule to distinguishbetween fast and slow motion frames Furthermore we use aconstant 120572 to adjust the decision rule

119865119895 119894119904 119891119886119904119905119890 119898119900119905119894119900119899 119891119903119886119898119890 119894119891 119872119872119872119881(119865119895) gt 119905ℎ119903119890119904ℎ119865119895 119894119904 119899119900 119898119900119905119894119900119899 119891119903119886119898119890 119894119891 119872119872119872119881(119865119895) le 119905ℎ119903119890119904ℎ (10)

Figure 3 shows the average motion energy of some frames inForeman Football and Akiyo video shots In the case of theForman video we see that the fast motion frames take placebetween the frames 150 and 228

32 Watermark Preprocessing In order to improve the secu-rity of the proposed algorithm the binary watermark shouldbe first preprocessed before embedded Here the watermark

is scrambled by using the chaotic logistic map which is deter-mined by the equation [25 27]119883119899+1 = 120582119883119899 (1 minus 119883119899) (11)

where 375 lt 120582 lt 4 is the system parameter The initial valueX0 L [0 1] is adopted as a key

Then the binary image logo or signature W(n) is scram-bled by X(n) with the following rule119882119875 (119899) = 119882 (119899) oplus 119883 (119899) 0 le 119899 le 119873 (12)

with N being the total number of bits in the watermark and oplusbeing the binary Exclusive Or (XOR) operation

33 Watermark Embedding Process In this section we de-scribe the proposed video watermarking scheme Figure 2shows the block diagram of the proposed video watermarkembedding procedure which is described as follows

(1) The fastmotion frames are extracted from the originalcolor video Only these frames are watermarked Thismakes the watermarked video quality good becausethe watermark is not embedded in all the frames as itis done in other watermarking schemes [6 9 28]

(2) Every fast motion frame is converted from the RGBto the YCbCr color space

(3) Every luminance frame is transformed with the 1-Level MR-SVD decomposition to get approximationand detail components 0 120593 1 1205932 1205933

(4) The approximation componentΦ is decomposed intoblocks of size [u times u]

(5) TheSVD is applied to each block of the approximationcomponent 0 (ie the low frequency subband) whichcontains the major video frame energy0119896119899 = 119880119896119899 times 119878119896119899 times 119881119879119896119899 (13)

where k is the fast motion frame index and n is thelocation of the block We use the SVD due to its sui-table properties discussed earlier

Security and Communication Networks 5

Watermark embedding process

Original video SVDOne level

MR-SVDutimesu

blocksRGB to YCbCr

FMFESBD

Watermark Chaotic Logistic Map Encryption

Watermarked video

YCbCr to RGB

Inverse MR-SVD

Compose utimesu blocks

Inverse SVD

３ＥＨ(11) =[ ３ＥＨ(11)

１]]１ +

3

4１ ７０(Ｈ) = 1

[ ３ＥＨ(11)１]１ +

1

4１ ７０(Ｈ) = 0

(a)

Watermark extraction process

Watermarked video

Recovered watermark

SBD FMFE SVDutimesu blocks

One level MR-SVD

RGB to YCbCr

Chaotic Logistic Map

decryption

７０

1 ３(11) minus １lfloor lfloor３

(11)

１ge

１

2

0 ３(11) minus １lfloor lfloor３

(11)

１lt

１

2

(b)

Figure 2 Proposed video watermarking technique (a) embedding process and (b) recover process

(6) The watermark is encrypted using chaotic logisticmap

(7) In order to guarantee robustness to our watermarkingscheme the watermarking system inserts the water-mark bits in the largest singular value using the QIMmethod as

119878119896119899 (1 1) = [119878119896119899 (1 1)119876 ]119876 + 34119876 119882119875 (119899) = 1[119878119896119899 (1 1)119876 ]119876 + 14119876 119882119875 (119899) = 0 (14)

where Q is the quantization step and [∙] stands for therounding operation

(8) Inverse SVD transformation is conducted to obtainthe watermarked block01015840119896119899 = 119880119896119899 times 1198781015840119896119899times119881119879119896119899 (15)

(9) To build the watermarked luminance inverse MR-SVD is applied to the modified approximation com-ponent 01015840

(10) Reconstruction of the watermarked video frame isdone using the watermarked luminance part and theoriginal frame chrominance parts Cb and Cr byconversion from the YCbCr into the RGB color space

34 Watermark Extraction Process The watermark extrac-tion process is shown in Figure 2 and is described as follows

(1) The watermarked fast motion frames are extractedfrom the watermarked color video

(2) Every watermarked fast motion frame is convertedfrom the RGB to the YCbCr

(3) Every watermarked luminance frame is transformedinto 1-LevelMR-SVD decomposition to get the water-marked approximation 01015840

(4) The watermarked approximation component is de-composed into blocks of size [u times u]

6 Security and Communication Networks

0

1

2

3

4

5

6

7The average motion energy in the second shot of video ForemanaviThe average motion energy in the first shot of video Foremanavi

220 240 260 280200Frame Number

20 40 60 80 100 120 140 160 1800Frame Number

0

1

2

3

4

5

6

7

MM

MV

MMMVThresh

MMMVThresh

(a)

4

5

The average motion energy in the second shot of video Fooballavi The average motion energy in the first shot of video Fooballavi

0

05

1

15

2

25

3

35

4

45

5

MM

MV

36

38

42

44

46

48

52

54

56

MM

MV

275 280 285 290 295270Frame Number

50 100 150 200 2500Frame Number

MMMVThresh

MMMVThresh

(b)

Average motion energy in Akiyoavi

0005

01015

02025

03035

MM

MV

50 100 150 200 250 3000Frame Number

MMMVThresh

(c)

Figure 3 Average motion energy in (a) Foreman (b) Football and (c) Akiyo videos

Security and Communication Networks 7

(5) TheSVD is applied to each block of the approximationcomponent01015840119896119899 = 119880119896119899 times 1198781015840119896119899times119881119879119896119899 (16)

(6) The embedded watermark is extracted by the follow-ing rule

119875 = 1 1198781015840 (1 1) minus 119876lfloor1198781015840 (1 1)119876 rfloor ge 11987620 1198781015840 (1 1) minus 119876lfloor1198781015840 (1 1)119876 rfloor lt 1198762 (17)

where lfloor∙rfloor is the floor function(7) Decryption with the same chaotic sequence is per-

formed to get the hidden binary watermark W

(8) Since a video clip contains several fast motion framesin which the same watermark is embedded we calcu-late the final recovered watermark by averaging thewatermarks extracted from these different frames

4 Results and Discussion

Theproposed algorithm is implemented using MATLABWeused three CIF (352times288) standard sequences shown in Fig-ure 4 (Foreman 300 frames Akiyo 300 frames and Football260 frames) The test videos are in RGB uncompressed aviformat with a frame rate of 30 fps [29] The watermark is abinary image Its resolution is 9 times 11 The block size in theproposed algorithm can be set to any desired value Howevera small block size leads to block effects problem whereas alarge block size reduces the total number of the watermarkbits We carried out various experiments and found that theblock size of 16x16 allows the embedding of an acceptablenumber of watermark bits without causing noticeable blockeffect

As shown inFigure 5 small values of the quantization stepyield good transparency at the expense of poor robustnessand vice versa From this figure it can be observed that thevalue of the quantization step Q = 110 gives a good com-promise between robustness and imperceptibility

41 Imperceptibility Tests The imperceptibility of the water-mark is estimated by measuring the PSNR (Peak Signal toNoise Ratio) and Mean Structure Similarity Index Measure(MSSIM) which are calculated using the luminance space Yof the original and watermarked frames [30] High PSNRvalues of the watermarked video frames indicate betterimperceptibility It is worth noting that the SSIM provides aperceptual distortion in range of [0 1] When both framesare numerically the same this value is equal to 1 The PSNRis calculated as follows

119875119878119873119877 = 10 log10 ( 2562119872119878119864) (18)

Table 1 The PSNR and MSSIM of the watermarked videos

Video Foreman Akiyo FootballAverage PSNR 4020 4045 4026Average MSSIM 09981 09988 09977

where the Mean Square Error (MSE) between the host lumi-nance Y and the watermarked luminance Y1015840 is defined as

119872119878119864 = 1119872 times119873119872minus1sum119894=0 119873minus1sum119895=0 1003816100381610038161003816119884 (119894 119895) minus 1198841015840 (119894 119895)100381610038161003816100381610038162 (19)

with M and N respectively being the height and width of thevideo frame

The MSSIM is defined as follows

119872119878119878119868119872(119884 1198841015840) = 1119872 119872sum119895=1 119878119878119868119872(119884119869 1198841015840119895) (20)

where YJ and Y1015840j are the image contents at the j local windowand M is the number of local windows of the image

119878119878119868119872(119884 1198841015840) = [119897 (119884 1198841015840)]120572 sdot [119888 (119884 1198841015840)]120573 sdot 119904 [119884 1198841015840)]120574 (21)

where l c and s are the luminance comparison the contrastcomparison and the structure comparison functions respec-tively With 120572 120573 and 120574 are parameters used to adjust therelative importance of the three components

As shown inTable 1 the average PSNRvalues of thewater-marked videos are higher than 40 dB and the correspondingMSSIMvalues are very close to 1This indicates the invisibilityof the watermark which means that the watermarked videosappear visually identical to the original ones as shown inFigure 6

42 Robustness Tests The robustness for any watermarkingsystem is a very important requirement To verify it weapply to the watermarked video various types of attacks andwe use the Normalized Coefficient (NC) and the Bit ErrorRate (BER) to compare the similarities between the originalwatermark W and the extracted watermark W The NC andBER are respectively calculated as

119873119862(119882 ) = sumsum 119882(119894 119895) (119894 119895)1003816100381610038161003816119882 (119894 119895)10038161003816100381610038162 (22)

119861119864119877 = 1119875 119901sum119895=1

10038161003816100381610038161003816 (119895) minus 119882 (119895)10038161003816100381610038161003816 (23)

where W W and P are respectively the original watermarkthe extracted watermark and the size of the watermark

The correlation between W and W is very high when NCis close to 1

Generally in video watermarking attacks are divided intothree categories image processing attacks frame synchro-nization attacks and video compression attacks

8 Security and Communication Networks

(a) (b)

(c)

Figure 4 Video sequences used for testing (a) Foreman (b) Akiyo and (c) Football

1

09965

0997

09975

0998

09985

0999

09995

MSS

IM

60 80 100 120 14040Q-step

(a)

60 80 100 120 14040Q-step

055

06

065

07

075

08

085

09

095

1

NC

(b)

Figure 5 (a) MSSIM of watermarked video for different quantization step Q and (b) NC of extracted watermark with JPEG compressionattack for different quantization step Q

Image Processing Attacks Considering a video as a sequenceof images the attacks applied to images can then be applied tothe video sequences The common image processing attacksare as follows

(i) Adding a noise three kinds of noises are added tothe watermarked video Gaussian noise salt amp peppernoise and speckle noise with density of 1 It canbe seen from Table 2 that the watermark is always

Security and Communication Networks 9

(a) (b)

(c)

Figure 6 Original and watermarked frames (a) frame 228 of Foreman (b) frame 16 of Akiyo and (c) frame 128 of Football

detectablewithNCandBER values respectively closeto 1 and 0 especially for Foreman and Football videosequences

(ii) Filtering 3x3median filter and 5x5 Gaussian filter areapplied separately to the watermarked videos and wecan see from Table 2 that the proposed method isrobust against median and Gaussian filtering

(iii) JPEG compression the watermarked video is com-pressed with different quality factors ranging from 10to 100 Figure 7 shows the results for the JPEG com-pression for instance if the watermarked video iscompressed with a quality factor of 40 the obtainedNC is greater than 95 and the BER value is lowerthan 02 This confirms the robustness of the pro-posed scheme to the JPEG compression attack

Frame Synchronization Attacks Because contents in the con-secutive frames of a video are almost identical it makesthe video sequences susceptible to temporal synchronizationattacks such as frame averaging frame dropping and frameswapping

(i) Frame dropping in frame dropping selected water-marked frames are replaced by their correspondingoriginal frames Table 3 shows the average NC andBER values given at different frame dropping ratesOur scheme achieves strong robustness against framedropping even for the case of high rates (ie 80)

(ii) Frame averaging in frame averaging we replaceselected watermarked frames by the average of theirprevious current and next frames The watermarkedvideo is averaged for various averaging rates and thenwe tried to extract the watermark Table 4 shows thatthe watermark can be recovered at frame averagingrates up to 50

(iii) Frame swapping the results presented in Table 5prove the robustness of our watermarking schemeagainst frame swapping because when all water-marked frames are swapped the NC is 1 and the BERis 0

Video Compression Attacks Video compression is a funda-mental attack in video watermarking that should be verifiedas video sequences are stored and transmitted in compressedformat Here we use a tool for video processing namedVirtualDub to compress the videos sequences with twodifferent Lossy compressions [31] H264 coding with a bit rateof 512kbps andMPEG4 coding with bit rates of 1500kbps and1000kbpsTheNC and BER are depicted in Table 6 and we seethat the algorithm can resist to video compression attacks

Results with the StirMark Benchmark We also evaluated theproposed method with StirMark 31 which is a well-knownevaluation tool for watermarking robustness of watermarkedvideo frames under image processing attacks [32 33] Table 7shows the evaluation of the proposed method with theStirMark benchmark under JPEG compression median filterwith JPEG compression Gaussian filter with JPEG compres-sion sharpening with JPEG compression and removal ofrow and clone with JPEG compression As can be observedthe proposed watermarking technique can resist to all thesenamed attacks

43 Comparison with Some Previously Reported AlgorithmsIn order to evaluate the performance of our algorithm wecompared the results of the proposed video watermarkingscheme to the results of related video watermarking schemesgiven in [6 16 18 20] which we introduced and discussedin the Introduction Figure 8 displays these results of com-parison under frame dropping frame averaging and JPEG

10 Security and Communication Networks

Table2Ex

tractedwatermark(w

ithNCandBE

Rvalues)u

nder

imagep

rocessingattack

Attacks

Foreman

Akiyo

Football

Noattack

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

Gaussianno

ise

NC=090

08BE

R=004

04NC=066

74BE

R=01414

NC=1

BER=0

SaltampPepp

erno

ise

NC=1

BER=0

NC=09238

BER=0303

NC=1

BER=0

Specklen

oise

NC=09479

BER=00202

NC=07870

BER=009

09NC=1

BER=0

Medianfilter

NC=09732

BER=0101

NC=1

BER=0

NC=1

BER=0

Gaussianlowpassfilter

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

Security and Communication Networks 11

Table3Ex

tractedwatermark(w

ithNCandBE

R)un

derframed

ropp

ingattack

Fram

edropp

ingrate

Foreman

Akiyo

Football

50

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

80

NC=08485

BER=006

06NC=07822

BER=0101

NC=07156

BER=01313

12 Security and Communication Networks

Table4Ex

tractedwatermark(w

ithNCandBE

R)un

derframea

veraging

attack

Fram

eaveraging

rate

Foreman

Akiyo

Football

25

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

50

NC=06819

BER=0111

1NC=0800

BER=009

09NC=1

BER=0

Security and Communication Networks 13

Table5Ex

tractedwatermark(w

ithNCandBE

R)un

derframes

wapping

attack

Fram

eSwapping

rate

Foreman

Akiyo

Football

50

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

100

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

14 Security and Communication Networks

Table6Ex

tractedwatermark(w

ithNCandBE

R)un

derv

ideo

compressio

nattack

Foreman

Akiyo

Football

H264compressio

n(512kbps)

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

MPE

G4compressio

n(1500kbp

s)NC=1

BER=0

NC=08708

BER=00505

NC=1

BER=0

MPE

G4compressio

n(100

0kbp

s)NC=1

BER=0

NC=0848

BER=006

06NC=09238

BER=00303

Security and Communication Networks 15

100

90

80

70

60

50

4010 20 30 40 50 60 70 80 90 100

NC

()

JPEG quality factor

10 20 30 40 50 60 70 80 90 100

JPEG quality factor

ForemanAkiyoFootball

ForemanAkiyoFootball

30

25

20

15

10

5

0

BER

()

Figure 7 NC and BER values of the extracted watermarks under the JPEG compression attack

Table 7 Extracted watermark (with NC and BER) under image processing attacks with StirMark benchmark

Attacks NC BERJPEG 10 05677 01515JPEG 30 09479 00202JPEG 60 08485 00606JPEG 90 10000 000003x3 median filter+ JPEG 09732 00101Gaussian filter +JPEG 07681 01010Sharpening +JPEG 05127 02626removed one row one colon +JPEG 05523 02020

compression attacks It is clear that the proposed schemeachieves a better robustness

44 Complexity of the Proposed Algorithm The proposedwatermarking schema has a low complexity because of thefollowing

(i) The algorithm is done in the MR-SVD domain whichhas an overall complexity O(n) where n indicates thesignal length [23]

(ii) To extract fast motion frames we use a simple thresh-old decision method based on motion energy

(iii) The embedding and extraction are not applied to allthe frames of the video but just to fast motion frames

In fact the execution time of the proposed scheme required inwatermark embedding and extraction processes the followingresults watermark embedding 036 s per frame watermarkextraction 019 s per frame The simulation is conducted on

Intel(R) Core i5 180 GHz processor with 4 GB RAMusingMATLAB verR2013b

5 Conclusion

In this paper we proposed a novel blind video watermarkingscheme in fast motion frames using the SVD the MR-SVD the QIM and the Logistic Map encryption In theproposed technique we solved the problem of embedding thewatermark in all frames by choosing only the frames that havebig motion energy to be the host frames suitable to the HSVUsing this approach we have limited the number of frames tobe processed and also assured a higher imperceptibility of thewatermark In addition the combination of the characteris-tics of the SVD MR-SVD QIM and Logistic Map Encryp-tion makes our scheme secure and robust to a variety ofattacks The experimental results confirm that the proposedwatermarking scheme has good imperceptibility with PSNRgreater than 40 dBTheproposed scheme is robust not only toimage processing attacks but also to frame synchronization

16 Security and Communication Networks

100

95

90

85

80

75

NC

()

NC

()

100

95

90

85

80

75

NC

()

10 15 20 25 30 35 40

frame averaging rate

proposed method[20][16][6]

proposed method

proposed method

[20]

[20]

[16][6]

[6]

[18]

10 20 30 40 50 60 70

frame dropping rate

70

65

100

95

90

85

80

75

70

65

60

JPEG Quality factor20 30 40 50 60 70 80 90

Figure 8 Comparison of experimental results between the proposed technique and other schemes under frame averaging attack framedropping attack and JPEG compression attack

and video compression attacks The comparison results withother algorithms related to video watermarking indicate thesuperiority of our scheme

Data Availability

The data used to support the findings of this study are avail-able from the corresponding author upon request

Conflicts of Interest

The authors declare no conflicts of interest

References

[1] J C IngemarM LMiller A B Jeffrey J Fridrich andT KalkerDigital Watermarking and Steganography Morgan KaufmannPublishers Inc 2008

[2] S Katzenbeisser and F Petitcolas ldquoInformation Hiding Tech-niques for Steganography and Digital Watermarkingrdquo ArtechHouse 2000

[3] H Tao L Chongmin J M Zain and A N Abdalla ldquoRobustimagewatermarking theories and techniques A reviewrdquo Journalof Applied Research and Technology vol 12 no 1 pp 122ndash1382014

Security and Communication Networks 17

[4] M Kutter F Jordan and F Bossen ldquoDigital watermarking ofcolor images using amplitude modulationrdquo Journal of ElectronicImaging vol 7 no 2 pp 326ndash332 1998

[5] A Karmakar A Phadikar B S Phadikar and G K Maity ldquoAblind video watermarking scheme resistant to rotation and col-lusion attacksrdquo Journal of King Saud University - Computer andInformation Sciences vol 28 no 2 pp 199ndash210 2016

[6] T R Singh K M Singh and S Roy ldquoVideo watermarkingscheme based on visual cryptography and scene change detec-tionrdquo AEU - International Journal of Electronics and Communi-cations vol 67 no 8 pp 645ndash651 2013

[7] W Kong B Yang D Wu and X Niu ldquoSVD Based Blind VideoWatermarking Algorithmrdquo in Proceedings of the First Interna-tional Conference on Innovative Computing Information andControl - Volume I (ICICICrsquo06) pp 265ndash268 Beijing ChinaSeptember 2006

[8] R Liu and T Tan ldquoAn SVD-based watermarking scheme forprotecting rightful ownershiprdquo IEEE Transactions on Multime-dia vol 4 no 1 pp 121ndash128 2002

[9] O S Faragallah ldquoEfficient video watermarking based on sin-gular value decomposition in the discrete wavelet transformdomainrdquo AEU - International Journal of Electronics and Com-munications vol 67 no 3 pp 189ndash196 2013

[10] K Niu X Yang and L Xiang ldquoHybrid quasi-3D DWTDCTand SVD video watermarkingrdquo in Proceedings of the 2010 IEEEInternational Conference on Software Engineering and ServiceSciences ICSESS 2010 pp 588ndash591 China July 2010

[11] N I Yassin NM Salem andM I El Adawy ldquoQIM blind videowatermarking scheme based on Wavelet transform and prin-cipal component analysisrdquo Alexandria Engineering Journal vol53 no 4 pp 833ndash842 2014

[12] D K Thind and S Jindal ldquoA semi blind DWT-SVD videowatermarkingrdquo inProceedings of the International Conference onInformation and Communication Technologies ICICT 2014 pp1661ndash1667 India December 2014

[13] N Leelavathy E V Prasad and S Kumar ldquoVideo Watermark-ing Techniques A Reviewrdquo International Journal of ComputerApplications vol 104 no 7 pp 24ndash30 2014

[14] F Hartung and M Kutter ldquoMultimedia watermarking techni-quesrdquo Proceedings of the IEEE vol 87 no 7 pp 1079ndash1107 1999

[15] P Rasti S Samiei M Agoyi S Escalera and G AnbarjafarildquoRobust non-blind color video watermarking using QR decom-position and entropy analysisrdquo Journal of Visual Communica-tion and Image Representation vol 38 pp 838ndash847 2016

[16] S M Youssef A A ElFarag and N M Ghatwary ldquoAdaptivevideo watermarking integrating a fuzzy wavelet-based humanvisual system perceptual modelrdquo Multimedia Tools and Appli-cations vol 73 no 3 pp 1545ndash1573 2014

[17] T Tabassum and S M Islam ldquoA digital video watermarkingtechnique based on identical frame extraction in 3-LevelDWTrdquoin Proceedings of the 2012 15th International Conference onComputer and Information Technology (ICCIT) pp 101ndash106Chittagong Bangladesh December 2012

[18] L Agilandeeswari and K Ganesan ldquoA robust color video water-marking scheme based on hybrid embedding techniquesrdquoMul-timedia Tools and Applications vol 75 no 14 pp 8745ndash87802016

[19] X Jiang Q Liu and Q Wu ldquoA new video watermarking algo-rithm based on shot segmentation and block classificationrdquoMultimedia Tools and Applications vol 62 no 3 pp 545ndash5602013

[20] K Chetan and K Raghavendra ldquoDWT Based Blind DigitalVideo Watermarking Scheme for Video Authenticationrdquo Inter-national Journal of Computer Applications vol 4 no 10 pp 19ndash26 2010

[21] Y JiaW Lin and A A Kassim ldquoEstimating just-noticeable dis-tortion for videordquo IEEE Transactions on Circuits and Systems forVideo Technology vol 16 no 7 pp 820ndash829 2006

[22] H C Andrews and C L Patterson ldquoSingular value decomposi-tions and digital image processingrdquo IEEE Transactions on SignalProcessing vol 24 no 1 pp 26ndash53 1976

[23] R Kakarala and P O Ogunbona ldquoSignal analysis using a multi-resolution form of the singular value decompositionrdquo IEEETransactions on Image Processing vol 10 no 5 pp 723ndash7352001

[24] B Chen and G WWornell ldquoQuantization index modulation aclass of provably goodmethods for digital watermarking and in-formation embeddingrdquo IEEE Transactions on Information The-ory vol 47 no 4 pp 1423ndash1443 2001

[25] M Yaghoobi ldquoA New Approach for Image Encryption UsingChaotic Logistic Maprdquo in Proceedings of the 2008 InternationalConference on Advanced Computer Theory and Engineering(ICACTE) pp 585ndash590 Phuket Thailand December 2008

[26] R Dugad K Ratakonda and N Ahuja ldquoRobust video shotchange detectionrdquo in Proceedings of the 1998 IEEE SecondWork-shop on Multimedia Signal Processing pp 376ndash381 RedondoBeach CA USA 1998

[27] A Anees I Hussain A Algarni and M Aslam ldquoA RobustWatermarking Scheme for Online Multimedia Copyright Pro-tection Using New Chaotic Maprdquo Security and CommunicationNetworks vol 2018 2018

[28] R Thanki V Dwivedi and K Borisagar ldquoA hybrid watermark-ing scheme with CS theory for security of multimedia datardquoJournal of King Saud University - Computer and InformationSciences 2017

[29] httpmediaxiphorgvideoderf Accessed on 06 July 2018[30] B G Haskell and A N Netravali Digital Pictures Representa-

tion Compression and Standards Perseus Publishing 1997[31] httpwwwvirtualduborg Accessed on 06 July 2018[32] F A Petitcolas R J Anderson and M G Kuhn ldquoAttacks on

CopyrightMarking Systemsrdquo in Information Hiding vol 1525 ofLecture Notes in Computer Science pp 218ndash238 Springer BerlinHeidelberg Berlin Heidelberg 1998

[33] httpwwwpetitcolasnetwatermarkingstirmark Accessed on19 September 2018

4 Security and Communication Networks

Figure 1 Original video frame Foreman image and its 1-level MR-SVD

the proposed system exploits the characteristics of the humanvisual system (HVS) to select frames in which the watermarkis embedded effectively Because the HVS is less sensitive toerrors in regions with great motion we select the frames thathave big motion energy to be the host frames [21] In ourscheme to extract the fast motion frames the cover video Vis first converted into individual frames then shot bound-aries (SBD) are detected using the algorithm proposed in[26] to obtain temporally stationary signals because frameswithin the same shot have a strong correlation Afterwardswe measure the average motion energy of each frame usingthe mean magnitudes of motion vectors (MMMV)

Let S = F1 F2 Fk be a shot of length k where Fk (k =1 2 k) represents the kth frame in the shot The MMMVof the frame can be calculated as follows119872119872119872119881(119865119895) = 1119873 sum

1le119894le1198731le119895le119896

1198721198812119909 (119894 119895) + 1198721198812119910 (119894 119895) (8)

where N is the number of macro blocks in a frame and MVxand MVy represent the components of the motion vector(MV) in respectively the X axis direction and the Y axisdirection

119879ℎ119903119890119904ℎ = 120572119896 119896sum119895=1 119872119872119872119881(119865119895)) (9)

The above threshold is used as a decision rule to distinguishbetween fast and slow motion frames Furthermore we use aconstant 120572 to adjust the decision rule

119865119895 119894119904 119891119886119904119905119890 119898119900119905119894119900119899 119891119903119886119898119890 119894119891 119872119872119872119881(119865119895) gt 119905ℎ119903119890119904ℎ119865119895 119894119904 119899119900 119898119900119905119894119900119899 119891119903119886119898119890 119894119891 119872119872119872119881(119865119895) le 119905ℎ119903119890119904ℎ (10)

Figure 3 shows the average motion energy of some frames inForeman Football and Akiyo video shots In the case of theForman video we see that the fast motion frames take placebetween the frames 150 and 228

32 Watermark Preprocessing In order to improve the secu-rity of the proposed algorithm the binary watermark shouldbe first preprocessed before embedded Here the watermark

is scrambled by using the chaotic logistic map which is deter-mined by the equation [25 27]119883119899+1 = 120582119883119899 (1 minus 119883119899) (11)

where 375 lt 120582 lt 4 is the system parameter The initial valueX0 L [0 1] is adopted as a key

Then the binary image logo or signature W(n) is scram-bled by X(n) with the following rule119882119875 (119899) = 119882 (119899) oplus 119883 (119899) 0 le 119899 le 119873 (12)

with N being the total number of bits in the watermark and oplusbeing the binary Exclusive Or (XOR) operation

33 Watermark Embedding Process In this section we de-scribe the proposed video watermarking scheme Figure 2shows the block diagram of the proposed video watermarkembedding procedure which is described as follows

(1) The fastmotion frames are extracted from the originalcolor video Only these frames are watermarked Thismakes the watermarked video quality good becausethe watermark is not embedded in all the frames as itis done in other watermarking schemes [6 9 28]

(2) Every fast motion frame is converted from the RGBto the YCbCr color space

(3) Every luminance frame is transformed with the 1-Level MR-SVD decomposition to get approximationand detail components 0 120593 1 1205932 1205933

(4) The approximation componentΦ is decomposed intoblocks of size [u times u]

(5) TheSVD is applied to each block of the approximationcomponent 0 (ie the low frequency subband) whichcontains the major video frame energy0119896119899 = 119880119896119899 times 119878119896119899 times 119881119879119896119899 (13)

where k is the fast motion frame index and n is thelocation of the block We use the SVD due to its sui-table properties discussed earlier

Security and Communication Networks 5

Watermark embedding process

Original video SVDOne level

MR-SVDutimesu

blocksRGB to YCbCr

FMFESBD

Watermark Chaotic Logistic Map Encryption

Watermarked video

YCbCr to RGB

Inverse MR-SVD

Compose utimesu blocks

Inverse SVD

３ＥＨ(11) =[ ３ＥＨ(11)

１]]１ +

3

4１ ７０(Ｈ) = 1

[ ３ＥＨ(11)１]１ +

1

4１ ７０(Ｈ) = 0

(a)

Watermark extraction process

Watermarked video

Recovered watermark

SBD FMFE SVDutimesu blocks

One level MR-SVD

RGB to YCbCr

Chaotic Logistic Map

decryption

７０

1 ３(11) minus １lfloor lfloor３

(11)

１ge

１

2

0 ３(11) minus １lfloor lfloor３

(11)

１lt

１

2

(b)

Figure 2 Proposed video watermarking technique (a) embedding process and (b) recover process

(6) The watermark is encrypted using chaotic logisticmap

(7) In order to guarantee robustness to our watermarkingscheme the watermarking system inserts the water-mark bits in the largest singular value using the QIMmethod as

119878119896119899 (1 1) = [119878119896119899 (1 1)119876 ]119876 + 34119876 119882119875 (119899) = 1[119878119896119899 (1 1)119876 ]119876 + 14119876 119882119875 (119899) = 0 (14)

where Q is the quantization step and [∙] stands for therounding operation

(8) Inverse SVD transformation is conducted to obtainthe watermarked block01015840119896119899 = 119880119896119899 times 1198781015840119896119899times119881119879119896119899 (15)

(9) To build the watermarked luminance inverse MR-SVD is applied to the modified approximation com-ponent 01015840

(10) Reconstruction of the watermarked video frame isdone using the watermarked luminance part and theoriginal frame chrominance parts Cb and Cr byconversion from the YCbCr into the RGB color space

34 Watermark Extraction Process The watermark extrac-tion process is shown in Figure 2 and is described as follows

(1) The watermarked fast motion frames are extractedfrom the watermarked color video

(2) Every watermarked fast motion frame is convertedfrom the RGB to the YCbCr

(3) Every watermarked luminance frame is transformedinto 1-LevelMR-SVD decomposition to get the water-marked approximation 01015840

(4) The watermarked approximation component is de-composed into blocks of size [u times u]

6 Security and Communication Networks

0

1

2

3

4

5

6

7The average motion energy in the second shot of video ForemanaviThe average motion energy in the first shot of video Foremanavi

220 240 260 280200Frame Number

20 40 60 80 100 120 140 160 1800Frame Number

0

1

2

3

4

5

6

7

MM

MV

MMMVThresh

MMMVThresh

(a)

4

5

The average motion energy in the second shot of video Fooballavi The average motion energy in the first shot of video Fooballavi

0

05

1

15

2

25

3

35

4

45

5

MM

MV

36

38

42

44

46

48

52

54

56

MM

MV

275 280 285 290 295270Frame Number

50 100 150 200 2500Frame Number

MMMVThresh

MMMVThresh

(b)

Average motion energy in Akiyoavi

0005

01015

02025

03035

MM

MV

50 100 150 200 250 3000Frame Number

MMMVThresh

(c)

Figure 3 Average motion energy in (a) Foreman (b) Football and (c) Akiyo videos

Security and Communication Networks 7

(5) TheSVD is applied to each block of the approximationcomponent01015840119896119899 = 119880119896119899 times 1198781015840119896119899times119881119879119896119899 (16)

(6) The embedded watermark is extracted by the follow-ing rule

119875 = 1 1198781015840 (1 1) minus 119876lfloor1198781015840 (1 1)119876 rfloor ge 11987620 1198781015840 (1 1) minus 119876lfloor1198781015840 (1 1)119876 rfloor lt 1198762 (17)

where lfloor∙rfloor is the floor function(7) Decryption with the same chaotic sequence is per-

formed to get the hidden binary watermark W

(8) Since a video clip contains several fast motion framesin which the same watermark is embedded we calcu-late the final recovered watermark by averaging thewatermarks extracted from these different frames

4 Results and Discussion

Theproposed algorithm is implemented using MATLABWeused three CIF (352times288) standard sequences shown in Fig-ure 4 (Foreman 300 frames Akiyo 300 frames and Football260 frames) The test videos are in RGB uncompressed aviformat with a frame rate of 30 fps [29] The watermark is abinary image Its resolution is 9 times 11 The block size in theproposed algorithm can be set to any desired value Howevera small block size leads to block effects problem whereas alarge block size reduces the total number of the watermarkbits We carried out various experiments and found that theblock size of 16x16 allows the embedding of an acceptablenumber of watermark bits without causing noticeable blockeffect

As shown inFigure 5 small values of the quantization stepyield good transparency at the expense of poor robustnessand vice versa From this figure it can be observed that thevalue of the quantization step Q = 110 gives a good com-promise between robustness and imperceptibility

41 Imperceptibility Tests The imperceptibility of the water-mark is estimated by measuring the PSNR (Peak Signal toNoise Ratio) and Mean Structure Similarity Index Measure(MSSIM) which are calculated using the luminance space Yof the original and watermarked frames [30] High PSNRvalues of the watermarked video frames indicate betterimperceptibility It is worth noting that the SSIM provides aperceptual distortion in range of [0 1] When both framesare numerically the same this value is equal to 1 The PSNRis calculated as follows

119875119878119873119877 = 10 log10 ( 2562119872119878119864) (18)

Table 1 The PSNR and MSSIM of the watermarked videos

Video Foreman Akiyo FootballAverage PSNR 4020 4045 4026Average MSSIM 09981 09988 09977

where the Mean Square Error (MSE) between the host lumi-nance Y and the watermarked luminance Y1015840 is defined as

119872119878119864 = 1119872 times119873119872minus1sum119894=0 119873minus1sum119895=0 1003816100381610038161003816119884 (119894 119895) minus 1198841015840 (119894 119895)100381610038161003816100381610038162 (19)

with M and N respectively being the height and width of thevideo frame

The MSSIM is defined as follows

119872119878119878119868119872(119884 1198841015840) = 1119872 119872sum119895=1 119878119878119868119872(119884119869 1198841015840119895) (20)

where YJ and Y1015840j are the image contents at the j local windowand M is the number of local windows of the image

119878119878119868119872(119884 1198841015840) = [119897 (119884 1198841015840)]120572 sdot [119888 (119884 1198841015840)]120573 sdot 119904 [119884 1198841015840)]120574 (21)

where l c and s are the luminance comparison the contrastcomparison and the structure comparison functions respec-tively With 120572 120573 and 120574 are parameters used to adjust therelative importance of the three components

As shown inTable 1 the average PSNRvalues of thewater-marked videos are higher than 40 dB and the correspondingMSSIMvalues are very close to 1This indicates the invisibilityof the watermark which means that the watermarked videosappear visually identical to the original ones as shown inFigure 6

42 Robustness Tests The robustness for any watermarkingsystem is a very important requirement To verify it weapply to the watermarked video various types of attacks andwe use the Normalized Coefficient (NC) and the Bit ErrorRate (BER) to compare the similarities between the originalwatermark W and the extracted watermark W The NC andBER are respectively calculated as

119873119862(119882 ) = sumsum 119882(119894 119895) (119894 119895)1003816100381610038161003816119882 (119894 119895)10038161003816100381610038162 (22)

119861119864119877 = 1119875 119901sum119895=1

10038161003816100381610038161003816 (119895) minus 119882 (119895)10038161003816100381610038161003816 (23)

where W W and P are respectively the original watermarkthe extracted watermark and the size of the watermark

The correlation between W and W is very high when NCis close to 1

Generally in video watermarking attacks are divided intothree categories image processing attacks frame synchro-nization attacks and video compression attacks

8 Security and Communication Networks

(a) (b)

(c)

Figure 4 Video sequences used for testing (a) Foreman (b) Akiyo and (c) Football

1

09965

0997

09975

0998

09985

0999

09995

MSS

IM

60 80 100 120 14040Q-step

(a)

60 80 100 120 14040Q-step

055

06

065

07

075

08

085

09

095

1

NC

(b)

Figure 5 (a) MSSIM of watermarked video for different quantization step Q and (b) NC of extracted watermark with JPEG compressionattack for different quantization step Q

Image Processing Attacks Considering a video as a sequenceof images the attacks applied to images can then be applied tothe video sequences The common image processing attacksare as follows

(i) Adding a noise three kinds of noises are added tothe watermarked video Gaussian noise salt amp peppernoise and speckle noise with density of 1 It canbe seen from Table 2 that the watermark is always

Security and Communication Networks 9

(a) (b)

(c)

Figure 6 Original and watermarked frames (a) frame 228 of Foreman (b) frame 16 of Akiyo and (c) frame 128 of Football

detectablewithNCandBER values respectively closeto 1 and 0 especially for Foreman and Football videosequences

(ii) Filtering 3x3median filter and 5x5 Gaussian filter areapplied separately to the watermarked videos and wecan see from Table 2 that the proposed method isrobust against median and Gaussian filtering

(iii) JPEG compression the watermarked video is com-pressed with different quality factors ranging from 10to 100 Figure 7 shows the results for the JPEG com-pression for instance if the watermarked video iscompressed with a quality factor of 40 the obtainedNC is greater than 95 and the BER value is lowerthan 02 This confirms the robustness of the pro-posed scheme to the JPEG compression attack

Frame Synchronization Attacks Because contents in the con-secutive frames of a video are almost identical it makesthe video sequences susceptible to temporal synchronizationattacks such as frame averaging frame dropping and frameswapping

(i) Frame dropping in frame dropping selected water-marked frames are replaced by their correspondingoriginal frames Table 3 shows the average NC andBER values given at different frame dropping ratesOur scheme achieves strong robustness against framedropping even for the case of high rates (ie 80)

(ii) Frame averaging in frame averaging we replaceselected watermarked frames by the average of theirprevious current and next frames The watermarkedvideo is averaged for various averaging rates and thenwe tried to extract the watermark Table 4 shows thatthe watermark can be recovered at frame averagingrates up to 50

(iii) Frame swapping the results presented in Table 5prove the robustness of our watermarking schemeagainst frame swapping because when all water-marked frames are swapped the NC is 1 and the BERis 0

Video Compression Attacks Video compression is a funda-mental attack in video watermarking that should be verifiedas video sequences are stored and transmitted in compressedformat Here we use a tool for video processing namedVirtualDub to compress the videos sequences with twodifferent Lossy compressions [31] H264 coding with a bit rateof 512kbps andMPEG4 coding with bit rates of 1500kbps and1000kbpsTheNC and BER are depicted in Table 6 and we seethat the algorithm can resist to video compression attacks

Results with the StirMark Benchmark We also evaluated theproposed method with StirMark 31 which is a well-knownevaluation tool for watermarking robustness of watermarkedvideo frames under image processing attacks [32 33] Table 7shows the evaluation of the proposed method with theStirMark benchmark under JPEG compression median filterwith JPEG compression Gaussian filter with JPEG compres-sion sharpening with JPEG compression and removal ofrow and clone with JPEG compression As can be observedthe proposed watermarking technique can resist to all thesenamed attacks

43 Comparison with Some Previously Reported AlgorithmsIn order to evaluate the performance of our algorithm wecompared the results of the proposed video watermarkingscheme to the results of related video watermarking schemesgiven in [6 16 18 20] which we introduced and discussedin the Introduction Figure 8 displays these results of com-parison under frame dropping frame averaging and JPEG

10 Security and Communication Networks

Table2Ex

tractedwatermark(w

ithNCandBE

Rvalues)u

nder

imagep

rocessingattack

Attacks

Foreman

Akiyo

Football

Noattack

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

Gaussianno

ise

NC=090

08BE

R=004

04NC=066

74BE

R=01414

NC=1

BER=0

SaltampPepp

erno

ise

NC=1

BER=0

NC=09238

BER=0303

NC=1

BER=0

Specklen

oise

NC=09479

BER=00202

NC=07870

BER=009

09NC=1

BER=0

Medianfilter

NC=09732

BER=0101

NC=1

BER=0

NC=1

BER=0

Gaussianlowpassfilter

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

Security and Communication Networks 11

Table3Ex

tractedwatermark(w

ithNCandBE

R)un

derframed

ropp

ingattack

Fram

edropp

ingrate

Foreman

Akiyo

Football

50

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

80

NC=08485

BER=006

06NC=07822

BER=0101

NC=07156

BER=01313

12 Security and Communication Networks

Table4Ex

tractedwatermark(w

ithNCandBE

R)un

derframea

veraging

attack

Fram

eaveraging

rate

Foreman

Akiyo

Football

25

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

50

NC=06819

BER=0111

1NC=0800

BER=009

09NC=1

BER=0

Security and Communication Networks 13

Table5Ex

tractedwatermark(w

ithNCandBE

R)un

derframes

wapping

attack

Fram

eSwapping

rate

Foreman

Akiyo

Football

50

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

100

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

14 Security and Communication Networks

Table6Ex

tractedwatermark(w

ithNCandBE

R)un

derv

ideo

compressio

nattack

Foreman

Akiyo

Football

H264compressio

n(512kbps)

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

MPE

G4compressio

n(1500kbp

s)NC=1

BER=0

NC=08708

BER=00505

NC=1

BER=0

MPE

G4compressio

n(100

0kbp

s)NC=1

BER=0

NC=0848

BER=006

06NC=09238

BER=00303

Security and Communication Networks 15

100

90

80

70

60

50

4010 20 30 40 50 60 70 80 90 100

NC

()

JPEG quality factor

10 20 30 40 50 60 70 80 90 100

JPEG quality factor

ForemanAkiyoFootball

ForemanAkiyoFootball

30

25

20

15

10

5

0

BER

()

Figure 7 NC and BER values of the extracted watermarks under the JPEG compression attack

Table 7 Extracted watermark (with NC and BER) under image processing attacks with StirMark benchmark

Attacks NC BERJPEG 10 05677 01515JPEG 30 09479 00202JPEG 60 08485 00606JPEG 90 10000 000003x3 median filter+ JPEG 09732 00101Gaussian filter +JPEG 07681 01010Sharpening +JPEG 05127 02626removed one row one colon +JPEG 05523 02020

compression attacks It is clear that the proposed schemeachieves a better robustness

44 Complexity of the Proposed Algorithm The proposedwatermarking schema has a low complexity because of thefollowing

(i) The algorithm is done in the MR-SVD domain whichhas an overall complexity O(n) where n indicates thesignal length [23]

(ii) To extract fast motion frames we use a simple thresh-old decision method based on motion energy

(iii) The embedding and extraction are not applied to allthe frames of the video but just to fast motion frames

In fact the execution time of the proposed scheme required inwatermark embedding and extraction processes the followingresults watermark embedding 036 s per frame watermarkextraction 019 s per frame The simulation is conducted on

Intel(R) Core i5 180 GHz processor with 4 GB RAMusingMATLAB verR2013b

5 Conclusion

In this paper we proposed a novel blind video watermarkingscheme in fast motion frames using the SVD the MR-SVD the QIM and the Logistic Map encryption In theproposed technique we solved the problem of embedding thewatermark in all frames by choosing only the frames that havebig motion energy to be the host frames suitable to the HSVUsing this approach we have limited the number of frames tobe processed and also assured a higher imperceptibility of thewatermark In addition the combination of the characteris-tics of the SVD MR-SVD QIM and Logistic Map Encryp-tion makes our scheme secure and robust to a variety ofattacks The experimental results confirm that the proposedwatermarking scheme has good imperceptibility with PSNRgreater than 40 dBTheproposed scheme is robust not only toimage processing attacks but also to frame synchronization

16 Security and Communication Networks

100

95

90

85

80

75

NC

()

NC

()

100

95

90

85

80

75

NC

()

10 15 20 25 30 35 40

frame averaging rate

proposed method[20][16][6]

proposed method

proposed method

[20]

[20]

[16][6]

[6]

[18]

10 20 30 40 50 60 70

frame dropping rate

70

65

100

95

90

85

80

75

70

65

60

JPEG Quality factor20 30 40 50 60 70 80 90

Figure 8 Comparison of experimental results between the proposed technique and other schemes under frame averaging attack framedropping attack and JPEG compression attack

and video compression attacks The comparison results withother algorithms related to video watermarking indicate thesuperiority of our scheme

Data Availability

The data used to support the findings of this study are avail-able from the corresponding author upon request

Conflicts of Interest

The authors declare no conflicts of interest

References

[1] J C IngemarM LMiller A B Jeffrey J Fridrich andT KalkerDigital Watermarking and Steganography Morgan KaufmannPublishers Inc 2008

[2] S Katzenbeisser and F Petitcolas ldquoInformation Hiding Tech-niques for Steganography and Digital Watermarkingrdquo ArtechHouse 2000

[3] H Tao L Chongmin J M Zain and A N Abdalla ldquoRobustimagewatermarking theories and techniques A reviewrdquo Journalof Applied Research and Technology vol 12 no 1 pp 122ndash1382014

Security and Communication Networks 17

[4] M Kutter F Jordan and F Bossen ldquoDigital watermarking ofcolor images using amplitude modulationrdquo Journal of ElectronicImaging vol 7 no 2 pp 326ndash332 1998

[5] A Karmakar A Phadikar B S Phadikar and G K Maity ldquoAblind video watermarking scheme resistant to rotation and col-lusion attacksrdquo Journal of King Saud University - Computer andInformation Sciences vol 28 no 2 pp 199ndash210 2016

[6] T R Singh K M Singh and S Roy ldquoVideo watermarkingscheme based on visual cryptography and scene change detec-tionrdquo AEU - International Journal of Electronics and Communi-cations vol 67 no 8 pp 645ndash651 2013

[7] W Kong B Yang D Wu and X Niu ldquoSVD Based Blind VideoWatermarking Algorithmrdquo in Proceedings of the First Interna-tional Conference on Innovative Computing Information andControl - Volume I (ICICICrsquo06) pp 265ndash268 Beijing ChinaSeptember 2006

[8] R Liu and T Tan ldquoAn SVD-based watermarking scheme forprotecting rightful ownershiprdquo IEEE Transactions on Multime-dia vol 4 no 1 pp 121ndash128 2002

[9] O S Faragallah ldquoEfficient video watermarking based on sin-gular value decomposition in the discrete wavelet transformdomainrdquo AEU - International Journal of Electronics and Com-munications vol 67 no 3 pp 189ndash196 2013

[10] K Niu X Yang and L Xiang ldquoHybrid quasi-3D DWTDCTand SVD video watermarkingrdquo in Proceedings of the 2010 IEEEInternational Conference on Software Engineering and ServiceSciences ICSESS 2010 pp 588ndash591 China July 2010

[11] N I Yassin NM Salem andM I El Adawy ldquoQIM blind videowatermarking scheme based on Wavelet transform and prin-cipal component analysisrdquo Alexandria Engineering Journal vol53 no 4 pp 833ndash842 2014

[12] D K Thind and S Jindal ldquoA semi blind DWT-SVD videowatermarkingrdquo inProceedings of the International Conference onInformation and Communication Technologies ICICT 2014 pp1661ndash1667 India December 2014

[13] N Leelavathy E V Prasad and S Kumar ldquoVideo Watermark-ing Techniques A Reviewrdquo International Journal of ComputerApplications vol 104 no 7 pp 24ndash30 2014

[14] F Hartung and M Kutter ldquoMultimedia watermarking techni-quesrdquo Proceedings of the IEEE vol 87 no 7 pp 1079ndash1107 1999

[15] P Rasti S Samiei M Agoyi S Escalera and G AnbarjafarildquoRobust non-blind color video watermarking using QR decom-position and entropy analysisrdquo Journal of Visual Communica-tion and Image Representation vol 38 pp 838ndash847 2016

[16] S M Youssef A A ElFarag and N M Ghatwary ldquoAdaptivevideo watermarking integrating a fuzzy wavelet-based humanvisual system perceptual modelrdquo Multimedia Tools and Appli-cations vol 73 no 3 pp 1545ndash1573 2014

[17] T Tabassum and S M Islam ldquoA digital video watermarkingtechnique based on identical frame extraction in 3-LevelDWTrdquoin Proceedings of the 2012 15th International Conference onComputer and Information Technology (ICCIT) pp 101ndash106Chittagong Bangladesh December 2012

[18] L Agilandeeswari and K Ganesan ldquoA robust color video water-marking scheme based on hybrid embedding techniquesrdquoMul-timedia Tools and Applications vol 75 no 14 pp 8745ndash87802016

[19] X Jiang Q Liu and Q Wu ldquoA new video watermarking algo-rithm based on shot segmentation and block classificationrdquoMultimedia Tools and Applications vol 62 no 3 pp 545ndash5602013

[20] K Chetan and K Raghavendra ldquoDWT Based Blind DigitalVideo Watermarking Scheme for Video Authenticationrdquo Inter-national Journal of Computer Applications vol 4 no 10 pp 19ndash26 2010

[21] Y JiaW Lin and A A Kassim ldquoEstimating just-noticeable dis-tortion for videordquo IEEE Transactions on Circuits and Systems forVideo Technology vol 16 no 7 pp 820ndash829 2006

[22] H C Andrews and C L Patterson ldquoSingular value decomposi-tions and digital image processingrdquo IEEE Transactions on SignalProcessing vol 24 no 1 pp 26ndash53 1976

[23] R Kakarala and P O Ogunbona ldquoSignal analysis using a multi-resolution form of the singular value decompositionrdquo IEEETransactions on Image Processing vol 10 no 5 pp 723ndash7352001

[24] B Chen and G WWornell ldquoQuantization index modulation aclass of provably goodmethods for digital watermarking and in-formation embeddingrdquo IEEE Transactions on Information The-ory vol 47 no 4 pp 1423ndash1443 2001

[25] M Yaghoobi ldquoA New Approach for Image Encryption UsingChaotic Logistic Maprdquo in Proceedings of the 2008 InternationalConference on Advanced Computer Theory and Engineering(ICACTE) pp 585ndash590 Phuket Thailand December 2008

[26] R Dugad K Ratakonda and N Ahuja ldquoRobust video shotchange detectionrdquo in Proceedings of the 1998 IEEE SecondWork-shop on Multimedia Signal Processing pp 376ndash381 RedondoBeach CA USA 1998

[27] A Anees I Hussain A Algarni and M Aslam ldquoA RobustWatermarking Scheme for Online Multimedia Copyright Pro-tection Using New Chaotic Maprdquo Security and CommunicationNetworks vol 2018 2018

[28] R Thanki V Dwivedi and K Borisagar ldquoA hybrid watermark-ing scheme with CS theory for security of multimedia datardquoJournal of King Saud University - Computer and InformationSciences 2017

[29] httpmediaxiphorgvideoderf Accessed on 06 July 2018[30] B G Haskell and A N Netravali Digital Pictures Representa-

tion Compression and Standards Perseus Publishing 1997[31] httpwwwvirtualduborg Accessed on 06 July 2018[32] F A Petitcolas R J Anderson and M G Kuhn ldquoAttacks on

CopyrightMarking Systemsrdquo in Information Hiding vol 1525 ofLecture Notes in Computer Science pp 218ndash238 Springer BerlinHeidelberg Berlin Heidelberg 1998

[33] httpwwwpetitcolasnetwatermarkingstirmark Accessed on19 September 2018

Security and Communication Networks 5

Watermark embedding process

Original video SVDOne level

MR-SVDutimesu

blocksRGB to YCbCr

FMFESBD

Watermark Chaotic Logistic Map Encryption

Watermarked video

YCbCr to RGB

Inverse MR-SVD

Compose utimesu blocks

Inverse SVD

３ＥＨ(11) =[ ３ＥＨ(11)

１]]１ +

3

4１ ７０(Ｈ) = 1

[ ３ＥＨ(11)１]１ +

1

4１ ７０(Ｈ) = 0

(a)

Watermark extraction process

Watermarked video

Recovered watermark

SBD FMFE SVDutimesu blocks

One level MR-SVD

RGB to YCbCr

Chaotic Logistic Map

decryption

７０

1 ３(11) minus １lfloor lfloor３

(11)

１ge

１

2

0 ３(11) minus １lfloor lfloor３

(11)

１lt

１

2

(b)

Figure 2 Proposed video watermarking technique (a) embedding process and (b) recover process

(6) The watermark is encrypted using chaotic logisticmap

(7) In order to guarantee robustness to our watermarkingscheme the watermarking system inserts the water-mark bits in the largest singular value using the QIMmethod as

119878119896119899 (1 1) = [119878119896119899 (1 1)119876 ]119876 + 34119876 119882119875 (119899) = 1[119878119896119899 (1 1)119876 ]119876 + 14119876 119882119875 (119899) = 0 (14)

where Q is the quantization step and [∙] stands for therounding operation

(8) Inverse SVD transformation is conducted to obtainthe watermarked block01015840119896119899 = 119880119896119899 times 1198781015840119896119899times119881119879119896119899 (15)

(9) To build the watermarked luminance inverse MR-SVD is applied to the modified approximation com-ponent 01015840

(10) Reconstruction of the watermarked video frame isdone using the watermarked luminance part and theoriginal frame chrominance parts Cb and Cr byconversion from the YCbCr into the RGB color space

34 Watermark Extraction Process The watermark extrac-tion process is shown in Figure 2 and is described as follows

(1) The watermarked fast motion frames are extractedfrom the watermarked color video

(2) Every watermarked fast motion frame is convertedfrom the RGB to the YCbCr

(3) Every watermarked luminance frame is transformedinto 1-LevelMR-SVD decomposition to get the water-marked approximation 01015840

(4) The watermarked approximation component is de-composed into blocks of size [u times u]

6 Security and Communication Networks

0

1

2

3

4

5

6

7The average motion energy in the second shot of video ForemanaviThe average motion energy in the first shot of video Foremanavi

220 240 260 280200Frame Number

20 40 60 80 100 120 140 160 1800Frame Number

0

1

2

3

4

5

6

7

MM

MV

MMMVThresh

MMMVThresh

(a)

4

5

The average motion energy in the second shot of video Fooballavi The average motion energy in the first shot of video Fooballavi

0

05

1

15

2

25

3

35

4

45

5

MM

MV

36

38

42

44

46

48

52

54

56

MM

MV

275 280 285 290 295270Frame Number

50 100 150 200 2500Frame Number

MMMVThresh

MMMVThresh

(b)

Average motion energy in Akiyoavi

0005

01015

02025

03035

MM

MV

50 100 150 200 250 3000Frame Number

MMMVThresh

(c)

Figure 3 Average motion energy in (a) Foreman (b) Football and (c) Akiyo videos

Security and Communication Networks 7

(5) TheSVD is applied to each block of the approximationcomponent01015840119896119899 = 119880119896119899 times 1198781015840119896119899times119881119879119896119899 (16)

(6) The embedded watermark is extracted by the follow-ing rule

119875 = 1 1198781015840 (1 1) minus 119876lfloor1198781015840 (1 1)119876 rfloor ge 11987620 1198781015840 (1 1) minus 119876lfloor1198781015840 (1 1)119876 rfloor lt 1198762 (17)

where lfloor∙rfloor is the floor function(7) Decryption with the same chaotic sequence is per-

formed to get the hidden binary watermark W

(8) Since a video clip contains several fast motion framesin which the same watermark is embedded we calcu-late the final recovered watermark by averaging thewatermarks extracted from these different frames

4 Results and Discussion

Theproposed algorithm is implemented using MATLABWeused three CIF (352times288) standard sequences shown in Fig-ure 4 (Foreman 300 frames Akiyo 300 frames and Football260 frames) The test videos are in RGB uncompressed aviformat with a frame rate of 30 fps [29] The watermark is abinary image Its resolution is 9 times 11 The block size in theproposed algorithm can be set to any desired value Howevera small block size leads to block effects problem whereas alarge block size reduces the total number of the watermarkbits We carried out various experiments and found that theblock size of 16x16 allows the embedding of an acceptablenumber of watermark bits without causing noticeable blockeffect

As shown inFigure 5 small values of the quantization stepyield good transparency at the expense of poor robustnessand vice versa From this figure it can be observed that thevalue of the quantization step Q = 110 gives a good com-promise between robustness and imperceptibility

41 Imperceptibility Tests The imperceptibility of the water-mark is estimated by measuring the PSNR (Peak Signal toNoise Ratio) and Mean Structure Similarity Index Measure(MSSIM) which are calculated using the luminance space Yof the original and watermarked frames [30] High PSNRvalues of the watermarked video frames indicate betterimperceptibility It is worth noting that the SSIM provides aperceptual distortion in range of [0 1] When both framesare numerically the same this value is equal to 1 The PSNRis calculated as follows

119875119878119873119877 = 10 log10 ( 2562119872119878119864) (18)

Table 1 The PSNR and MSSIM of the watermarked videos

Video Foreman Akiyo FootballAverage PSNR 4020 4045 4026Average MSSIM 09981 09988 09977

where the Mean Square Error (MSE) between the host lumi-nance Y and the watermarked luminance Y1015840 is defined as

119872119878119864 = 1119872 times119873119872minus1sum119894=0 119873minus1sum119895=0 1003816100381610038161003816119884 (119894 119895) minus 1198841015840 (119894 119895)100381610038161003816100381610038162 (19)

with M and N respectively being the height and width of thevideo frame

The MSSIM is defined as follows

119872119878119878119868119872(119884 1198841015840) = 1119872 119872sum119895=1 119878119878119868119872(119884119869 1198841015840119895) (20)

where YJ and Y1015840j are the image contents at the j local windowand M is the number of local windows of the image

119878119878119868119872(119884 1198841015840) = [119897 (119884 1198841015840)]120572 sdot [119888 (119884 1198841015840)]120573 sdot 119904 [119884 1198841015840)]120574 (21)

where l c and s are the luminance comparison the contrastcomparison and the structure comparison functions respec-tively With 120572 120573 and 120574 are parameters used to adjust therelative importance of the three components

As shown inTable 1 the average PSNRvalues of thewater-marked videos are higher than 40 dB and the correspondingMSSIMvalues are very close to 1This indicates the invisibilityof the watermark which means that the watermarked videosappear visually identical to the original ones as shown inFigure 6

42 Robustness Tests The robustness for any watermarkingsystem is a very important requirement To verify it weapply to the watermarked video various types of attacks andwe use the Normalized Coefficient (NC) and the Bit ErrorRate (BER) to compare the similarities between the originalwatermark W and the extracted watermark W The NC andBER are respectively calculated as

119873119862(119882 ) = sumsum 119882(119894 119895) (119894 119895)1003816100381610038161003816119882 (119894 119895)10038161003816100381610038162 (22)

119861119864119877 = 1119875 119901sum119895=1

10038161003816100381610038161003816 (119895) minus 119882 (119895)10038161003816100381610038161003816 (23)

where W W and P are respectively the original watermarkthe extracted watermark and the size of the watermark

The correlation between W and W is very high when NCis close to 1

Generally in video watermarking attacks are divided intothree categories image processing attacks frame synchro-nization attacks and video compression attacks

8 Security and Communication Networks

(a) (b)

(c)

Figure 4 Video sequences used for testing (a) Foreman (b) Akiyo and (c) Football

1

09965

0997

09975

0998

09985

0999

09995

MSS

IM

60 80 100 120 14040Q-step

(a)

60 80 100 120 14040Q-step

055

06

065

07

075

08

085

09

095

1

NC

(b)

Figure 5 (a) MSSIM of watermarked video for different quantization step Q and (b) NC of extracted watermark with JPEG compressionattack for different quantization step Q

Image Processing Attacks Considering a video as a sequenceof images the attacks applied to images can then be applied tothe video sequences The common image processing attacksare as follows

(i) Adding a noise three kinds of noises are added tothe watermarked video Gaussian noise salt amp peppernoise and speckle noise with density of 1 It canbe seen from Table 2 that the watermark is always

Security and Communication Networks 9

(a) (b)

(c)

Figure 6 Original and watermarked frames (a) frame 228 of Foreman (b) frame 16 of Akiyo and (c) frame 128 of Football

detectablewithNCandBER values respectively closeto 1 and 0 especially for Foreman and Football videosequences

(ii) Filtering 3x3median filter and 5x5 Gaussian filter areapplied separately to the watermarked videos and wecan see from Table 2 that the proposed method isrobust against median and Gaussian filtering

(iii) JPEG compression the watermarked video is com-pressed with different quality factors ranging from 10to 100 Figure 7 shows the results for the JPEG com-pression for instance if the watermarked video iscompressed with a quality factor of 40 the obtainedNC is greater than 95 and the BER value is lowerthan 02 This confirms the robustness of the pro-posed scheme to the JPEG compression attack

Frame Synchronization Attacks Because contents in the con-secutive frames of a video are almost identical it makesthe video sequences susceptible to temporal synchronizationattacks such as frame averaging frame dropping and frameswapping

(i) Frame dropping in frame dropping selected water-marked frames are replaced by their correspondingoriginal frames Table 3 shows the average NC andBER values given at different frame dropping ratesOur scheme achieves strong robustness against framedropping even for the case of high rates (ie 80)

(ii) Frame averaging in frame averaging we replaceselected watermarked frames by the average of theirprevious current and next frames The watermarkedvideo is averaged for various averaging rates and thenwe tried to extract the watermark Table 4 shows thatthe watermark can be recovered at frame averagingrates up to 50

(iii) Frame swapping the results presented in Table 5prove the robustness of our watermarking schemeagainst frame swapping because when all water-marked frames are swapped the NC is 1 and the BERis 0

Video Compression Attacks Video compression is a funda-mental attack in video watermarking that should be verifiedas video sequences are stored and transmitted in compressedformat Here we use a tool for video processing namedVirtualDub to compress the videos sequences with twodifferent Lossy compressions [31] H264 coding with a bit rateof 512kbps andMPEG4 coding with bit rates of 1500kbps and1000kbpsTheNC and BER are depicted in Table 6 and we seethat the algorithm can resist to video compression attacks

Results with the StirMark Benchmark We also evaluated theproposed method with StirMark 31 which is a well-knownevaluation tool for watermarking robustness of watermarkedvideo frames under image processing attacks [32 33] Table 7shows the evaluation of the proposed method with theStirMark benchmark under JPEG compression median filterwith JPEG compression Gaussian filter with JPEG compres-sion sharpening with JPEG compression and removal ofrow and clone with JPEG compression As can be observedthe proposed watermarking technique can resist to all thesenamed attacks

43 Comparison with Some Previously Reported AlgorithmsIn order to evaluate the performance of our algorithm wecompared the results of the proposed video watermarkingscheme to the results of related video watermarking schemesgiven in [6 16 18 20] which we introduced and discussedin the Introduction Figure 8 displays these results of com-parison under frame dropping frame averaging and JPEG

10 Security and Communication Networks

Table2Ex

tractedwatermark(w

ithNCandBE

Rvalues)u

nder

imagep

rocessingattack

Attacks

Foreman

Akiyo

Football

Noattack

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

Gaussianno

ise

NC=090

08BE

R=004

04NC=066

74BE

R=01414

NC=1

BER=0

SaltampPepp

erno

ise

NC=1

BER=0

NC=09238

BER=0303

NC=1

BER=0

Specklen

oise

NC=09479

BER=00202

NC=07870

BER=009

09NC=1

BER=0

Medianfilter

NC=09732

BER=0101

NC=1

BER=0

NC=1

BER=0

Gaussianlowpassfilter

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

Security and Communication Networks 11

Table3Ex

tractedwatermark(w

ithNCandBE

R)un

derframed

ropp

ingattack

Fram

edropp

ingrate

Foreman

Akiyo

Football

50

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

80

NC=08485

BER=006

06NC=07822

BER=0101

NC=07156

BER=01313

12 Security and Communication Networks

Table4Ex

tractedwatermark(w

ithNCandBE

R)un

derframea

veraging

attack

Fram

eaveraging

rate

Foreman

Akiyo

Football

25

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

50

NC=06819

BER=0111

1NC=0800

BER=009

09NC=1

BER=0

Security and Communication Networks 13

Table5Ex

tractedwatermark(w

ithNCandBE

R)un

derframes

wapping

attack

Fram

eSwapping

rate

Foreman

Akiyo

Football

50

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

100

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

14 Security and Communication Networks

Table6Ex

tractedwatermark(w

ithNCandBE

R)un

derv

ideo

compressio

nattack

Foreman

Akiyo

Football

H264compressio

n(512kbps)

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

MPE

G4compressio

n(1500kbp

s)NC=1

BER=0

NC=08708

BER=00505

NC=1

BER=0

MPE

G4compressio

n(100

0kbp

s)NC=1

BER=0

NC=0848

BER=006

06NC=09238

BER=00303

Security and Communication Networks 15

100

90

80

70

60

50

4010 20 30 40 50 60 70 80 90 100

NC

()

JPEG quality factor

10 20 30 40 50 60 70 80 90 100

JPEG quality factor

ForemanAkiyoFootball

ForemanAkiyoFootball

30

25

20

15

10

5

0

BER

()

Figure 7 NC and BER values of the extracted watermarks under the JPEG compression attack

Table 7 Extracted watermark (with NC and BER) under image processing attacks with StirMark benchmark

Attacks NC BERJPEG 10 05677 01515JPEG 30 09479 00202JPEG 60 08485 00606JPEG 90 10000 000003x3 median filter+ JPEG 09732 00101Gaussian filter +JPEG 07681 01010Sharpening +JPEG 05127 02626removed one row one colon +JPEG 05523 02020

compression attacks It is clear that the proposed schemeachieves a better robustness

44 Complexity of the Proposed Algorithm The proposedwatermarking schema has a low complexity because of thefollowing

(i) The algorithm is done in the MR-SVD domain whichhas an overall complexity O(n) where n indicates thesignal length [23]

(ii) To extract fast motion frames we use a simple thresh-old decision method based on motion energy

(iii) The embedding and extraction are not applied to allthe frames of the video but just to fast motion frames

In fact the execution time of the proposed scheme required inwatermark embedding and extraction processes the followingresults watermark embedding 036 s per frame watermarkextraction 019 s per frame The simulation is conducted on

Intel(R) Core i5 180 GHz processor with 4 GB RAMusingMATLAB verR2013b

5 Conclusion

In this paper we proposed a novel blind video watermarkingscheme in fast motion frames using the SVD the MR-SVD the QIM and the Logistic Map encryption In theproposed technique we solved the problem of embedding thewatermark in all frames by choosing only the frames that havebig motion energy to be the host frames suitable to the HSVUsing this approach we have limited the number of frames tobe processed and also assured a higher imperceptibility of thewatermark In addition the combination of the characteris-tics of the SVD MR-SVD QIM and Logistic Map Encryp-tion makes our scheme secure and robust to a variety ofattacks The experimental results confirm that the proposedwatermarking scheme has good imperceptibility with PSNRgreater than 40 dBTheproposed scheme is robust not only toimage processing attacks but also to frame synchronization

16 Security and Communication Networks

100

95

90

85

80

75

NC

()

NC

()

100

95

90

85

80

75

NC

()

10 15 20 25 30 35 40

frame averaging rate

proposed method[20][16][6]

proposed method

proposed method

[20]

[20]

[16][6]

[6]

[18]

10 20 30 40 50 60 70

frame dropping rate

70

65

100

95

90

85

80

75

70

65

60

JPEG Quality factor20 30 40 50 60 70 80 90

Figure 8 Comparison of experimental results between the proposed technique and other schemes under frame averaging attack framedropping attack and JPEG compression attack

and video compression attacks The comparison results withother algorithms related to video watermarking indicate thesuperiority of our scheme

Data Availability

The data used to support the findings of this study are avail-able from the corresponding author upon request

Conflicts of Interest

The authors declare no conflicts of interest

References

[1] J C IngemarM LMiller A B Jeffrey J Fridrich andT KalkerDigital Watermarking and Steganography Morgan KaufmannPublishers Inc 2008

[2] S Katzenbeisser and F Petitcolas ldquoInformation Hiding Tech-niques for Steganography and Digital Watermarkingrdquo ArtechHouse 2000

[3] H Tao L Chongmin J M Zain and A N Abdalla ldquoRobustimagewatermarking theories and techniques A reviewrdquo Journalof Applied Research and Technology vol 12 no 1 pp 122ndash1382014

Security and Communication Networks 17

[4] M Kutter F Jordan and F Bossen ldquoDigital watermarking ofcolor images using amplitude modulationrdquo Journal of ElectronicImaging vol 7 no 2 pp 326ndash332 1998

[5] A Karmakar A Phadikar B S Phadikar and G K Maity ldquoAblind video watermarking scheme resistant to rotation and col-lusion attacksrdquo Journal of King Saud University - Computer andInformation Sciences vol 28 no 2 pp 199ndash210 2016

[6] T R Singh K M Singh and S Roy ldquoVideo watermarkingscheme based on visual cryptography and scene change detec-tionrdquo AEU - International Journal of Electronics and Communi-cations vol 67 no 8 pp 645ndash651 2013

[7] W Kong B Yang D Wu and X Niu ldquoSVD Based Blind VideoWatermarking Algorithmrdquo in Proceedings of the First Interna-tional Conference on Innovative Computing Information andControl - Volume I (ICICICrsquo06) pp 265ndash268 Beijing ChinaSeptember 2006

[8] R Liu and T Tan ldquoAn SVD-based watermarking scheme forprotecting rightful ownershiprdquo IEEE Transactions on Multime-dia vol 4 no 1 pp 121ndash128 2002

[9] O S Faragallah ldquoEfficient video watermarking based on sin-gular value decomposition in the discrete wavelet transformdomainrdquo AEU - International Journal of Electronics and Com-munications vol 67 no 3 pp 189ndash196 2013

[10] K Niu X Yang and L Xiang ldquoHybrid quasi-3D DWTDCTand SVD video watermarkingrdquo in Proceedings of the 2010 IEEEInternational Conference on Software Engineering and ServiceSciences ICSESS 2010 pp 588ndash591 China July 2010

[11] N I Yassin NM Salem andM I El Adawy ldquoQIM blind videowatermarking scheme based on Wavelet transform and prin-cipal component analysisrdquo Alexandria Engineering Journal vol53 no 4 pp 833ndash842 2014

[12] D K Thind and S Jindal ldquoA semi blind DWT-SVD videowatermarkingrdquo inProceedings of the International Conference onInformation and Communication Technologies ICICT 2014 pp1661ndash1667 India December 2014

[13] N Leelavathy E V Prasad and S Kumar ldquoVideo Watermark-ing Techniques A Reviewrdquo International Journal of ComputerApplications vol 104 no 7 pp 24ndash30 2014

[14] F Hartung and M Kutter ldquoMultimedia watermarking techni-quesrdquo Proceedings of the IEEE vol 87 no 7 pp 1079ndash1107 1999

[15] P Rasti S Samiei M Agoyi S Escalera and G AnbarjafarildquoRobust non-blind color video watermarking using QR decom-position and entropy analysisrdquo Journal of Visual Communica-tion and Image Representation vol 38 pp 838ndash847 2016

[16] S M Youssef A A ElFarag and N M Ghatwary ldquoAdaptivevideo watermarking integrating a fuzzy wavelet-based humanvisual system perceptual modelrdquo Multimedia Tools and Appli-cations vol 73 no 3 pp 1545ndash1573 2014

[17] T Tabassum and S M Islam ldquoA digital video watermarkingtechnique based on identical frame extraction in 3-LevelDWTrdquoin Proceedings of the 2012 15th International Conference onComputer and Information Technology (ICCIT) pp 101ndash106Chittagong Bangladesh December 2012

[18] L Agilandeeswari and K Ganesan ldquoA robust color video water-marking scheme based on hybrid embedding techniquesrdquoMul-timedia Tools and Applications vol 75 no 14 pp 8745ndash87802016

[19] X Jiang Q Liu and Q Wu ldquoA new video watermarking algo-rithm based on shot segmentation and block classificationrdquoMultimedia Tools and Applications vol 62 no 3 pp 545ndash5602013

[20] K Chetan and K Raghavendra ldquoDWT Based Blind DigitalVideo Watermarking Scheme for Video Authenticationrdquo Inter-national Journal of Computer Applications vol 4 no 10 pp 19ndash26 2010

[21] Y JiaW Lin and A A Kassim ldquoEstimating just-noticeable dis-tortion for videordquo IEEE Transactions on Circuits and Systems forVideo Technology vol 16 no 7 pp 820ndash829 2006

[22] H C Andrews and C L Patterson ldquoSingular value decomposi-tions and digital image processingrdquo IEEE Transactions on SignalProcessing vol 24 no 1 pp 26ndash53 1976

[23] R Kakarala and P O Ogunbona ldquoSignal analysis using a multi-resolution form of the singular value decompositionrdquo IEEETransactions on Image Processing vol 10 no 5 pp 723ndash7352001

[24] B Chen and G WWornell ldquoQuantization index modulation aclass of provably goodmethods for digital watermarking and in-formation embeddingrdquo IEEE Transactions on Information The-ory vol 47 no 4 pp 1423ndash1443 2001

[25] M Yaghoobi ldquoA New Approach for Image Encryption UsingChaotic Logistic Maprdquo in Proceedings of the 2008 InternationalConference on Advanced Computer Theory and Engineering(ICACTE) pp 585ndash590 Phuket Thailand December 2008

[26] R Dugad K Ratakonda and N Ahuja ldquoRobust video shotchange detectionrdquo in Proceedings of the 1998 IEEE SecondWork-shop on Multimedia Signal Processing pp 376ndash381 RedondoBeach CA USA 1998

[27] A Anees I Hussain A Algarni and M Aslam ldquoA RobustWatermarking Scheme for Online Multimedia Copyright Pro-tection Using New Chaotic Maprdquo Security and CommunicationNetworks vol 2018 2018

[28] R Thanki V Dwivedi and K Borisagar ldquoA hybrid watermark-ing scheme with CS theory for security of multimedia datardquoJournal of King Saud University - Computer and InformationSciences 2017

[29] httpmediaxiphorgvideoderf Accessed on 06 July 2018[30] B G Haskell and A N Netravali Digital Pictures Representa-

tion Compression and Standards Perseus Publishing 1997[31] httpwwwvirtualduborg Accessed on 06 July 2018[32] F A Petitcolas R J Anderson and M G Kuhn ldquoAttacks on

CopyrightMarking Systemsrdquo in Information Hiding vol 1525 ofLecture Notes in Computer Science pp 218ndash238 Springer BerlinHeidelberg Berlin Heidelberg 1998

[33] httpwwwpetitcolasnetwatermarkingstirmark Accessed on19 September 2018

6 Security and Communication Networks

0

1

2

3

4

5

6

7The average motion energy in the second shot of video ForemanaviThe average motion energy in the first shot of video Foremanavi

220 240 260 280200Frame Number

20 40 60 80 100 120 140 160 1800Frame Number

0

1

2

3

4

5

6

7

MM

MV

MMMVThresh

MMMVThresh

(a)

4

5

The average motion energy in the second shot of video Fooballavi The average motion energy in the first shot of video Fooballavi

0

05

1

15

2

25

3

35

4

45

5

MM

MV

36

38

42

44

46

48

52

54

56

MM

MV

275 280 285 290 295270Frame Number

50 100 150 200 2500Frame Number

MMMVThresh

MMMVThresh

(b)

Average motion energy in Akiyoavi

0005

01015

02025

03035

MM

MV

50 100 150 200 250 3000Frame Number

MMMVThresh

(c)

Figure 3 Average motion energy in (a) Foreman (b) Football and (c) Akiyo videos

Security and Communication Networks 7

(5) TheSVD is applied to each block of the approximationcomponent01015840119896119899 = 119880119896119899 times 1198781015840119896119899times119881119879119896119899 (16)

(6) The embedded watermark is extracted by the follow-ing rule

119875 = 1 1198781015840 (1 1) minus 119876lfloor1198781015840 (1 1)119876 rfloor ge 11987620 1198781015840 (1 1) minus 119876lfloor1198781015840 (1 1)119876 rfloor lt 1198762 (17)

where lfloor∙rfloor is the floor function(7) Decryption with the same chaotic sequence is per-

formed to get the hidden binary watermark W

(8) Since a video clip contains several fast motion framesin which the same watermark is embedded we calcu-late the final recovered watermark by averaging thewatermarks extracted from these different frames

4 Results and Discussion

Theproposed algorithm is implemented using MATLABWeused three CIF (352times288) standard sequences shown in Fig-ure 4 (Foreman 300 frames Akiyo 300 frames and Football260 frames) The test videos are in RGB uncompressed aviformat with a frame rate of 30 fps [29] The watermark is abinary image Its resolution is 9 times 11 The block size in theproposed algorithm can be set to any desired value Howevera small block size leads to block effects problem whereas alarge block size reduces the total number of the watermarkbits We carried out various experiments and found that theblock size of 16x16 allows the embedding of an acceptablenumber of watermark bits without causing noticeable blockeffect

As shown inFigure 5 small values of the quantization stepyield good transparency at the expense of poor robustnessand vice versa From this figure it can be observed that thevalue of the quantization step Q = 110 gives a good com-promise between robustness and imperceptibility

41 Imperceptibility Tests The imperceptibility of the water-mark is estimated by measuring the PSNR (Peak Signal toNoise Ratio) and Mean Structure Similarity Index Measure(MSSIM) which are calculated using the luminance space Yof the original and watermarked frames [30] High PSNRvalues of the watermarked video frames indicate betterimperceptibility It is worth noting that the SSIM provides aperceptual distortion in range of [0 1] When both framesare numerically the same this value is equal to 1 The PSNRis calculated as follows

119875119878119873119877 = 10 log10 ( 2562119872119878119864) (18)

Table 1 The PSNR and MSSIM of the watermarked videos

Video Foreman Akiyo FootballAverage PSNR 4020 4045 4026Average MSSIM 09981 09988 09977

where the Mean Square Error (MSE) between the host lumi-nance Y and the watermarked luminance Y1015840 is defined as

119872119878119864 = 1119872 times119873119872minus1sum119894=0 119873minus1sum119895=0 1003816100381610038161003816119884 (119894 119895) minus 1198841015840 (119894 119895)100381610038161003816100381610038162 (19)

with M and N respectively being the height and width of thevideo frame

The MSSIM is defined as follows

119872119878119878119868119872(119884 1198841015840) = 1119872 119872sum119895=1 119878119878119868119872(119884119869 1198841015840119895) (20)

where YJ and Y1015840j are the image contents at the j local windowand M is the number of local windows of the image

119878119878119868119872(119884 1198841015840) = [119897 (119884 1198841015840)]120572 sdot [119888 (119884 1198841015840)]120573 sdot 119904 [119884 1198841015840)]120574 (21)

where l c and s are the luminance comparison the contrastcomparison and the structure comparison functions respec-tively With 120572 120573 and 120574 are parameters used to adjust therelative importance of the three components

As shown inTable 1 the average PSNRvalues of thewater-marked videos are higher than 40 dB and the correspondingMSSIMvalues are very close to 1This indicates the invisibilityof the watermark which means that the watermarked videosappear visually identical to the original ones as shown inFigure 6

42 Robustness Tests The robustness for any watermarkingsystem is a very important requirement To verify it weapply to the watermarked video various types of attacks andwe use the Normalized Coefficient (NC) and the Bit ErrorRate (BER) to compare the similarities between the originalwatermark W and the extracted watermark W The NC andBER are respectively calculated as

119873119862(119882 ) = sumsum 119882(119894 119895) (119894 119895)1003816100381610038161003816119882 (119894 119895)10038161003816100381610038162 (22)

119861119864119877 = 1119875 119901sum119895=1

10038161003816100381610038161003816 (119895) minus 119882 (119895)10038161003816100381610038161003816 (23)

where W W and P are respectively the original watermarkthe extracted watermark and the size of the watermark

The correlation between W and W is very high when NCis close to 1

Generally in video watermarking attacks are divided intothree categories image processing attacks frame synchro-nization attacks and video compression attacks

8 Security and Communication Networks

(a) (b)

(c)

Figure 4 Video sequences used for testing (a) Foreman (b) Akiyo and (c) Football

1

09965

0997

09975

0998

09985

0999

09995

MSS

IM

60 80 100 120 14040Q-step

(a)

60 80 100 120 14040Q-step

055

06

065

07

075

08

085

09

095

1

NC

(b)

Figure 5 (a) MSSIM of watermarked video for different quantization step Q and (b) NC of extracted watermark with JPEG compressionattack for different quantization step Q

Image Processing Attacks Considering a video as a sequenceof images the attacks applied to images can then be applied tothe video sequences The common image processing attacksare as follows

(i) Adding a noise three kinds of noises are added tothe watermarked video Gaussian noise salt amp peppernoise and speckle noise with density of 1 It canbe seen from Table 2 that the watermark is always

Security and Communication Networks 9

(a) (b)

(c)

Figure 6 Original and watermarked frames (a) frame 228 of Foreman (b) frame 16 of Akiyo and (c) frame 128 of Football

detectablewithNCandBER values respectively closeto 1 and 0 especially for Foreman and Football videosequences

(ii) Filtering 3x3median filter and 5x5 Gaussian filter areapplied separately to the watermarked videos and wecan see from Table 2 that the proposed method isrobust against median and Gaussian filtering

(iii) JPEG compression the watermarked video is com-pressed with different quality factors ranging from 10to 100 Figure 7 shows the results for the JPEG com-pression for instance if the watermarked video iscompressed with a quality factor of 40 the obtainedNC is greater than 95 and the BER value is lowerthan 02 This confirms the robustness of the pro-posed scheme to the JPEG compression attack

Frame Synchronization Attacks Because contents in the con-secutive frames of a video are almost identical it makesthe video sequences susceptible to temporal synchronizationattacks such as frame averaging frame dropping and frameswapping

(i) Frame dropping in frame dropping selected water-marked frames are replaced by their correspondingoriginal frames Table 3 shows the average NC andBER values given at different frame dropping ratesOur scheme achieves strong robustness against framedropping even for the case of high rates (ie 80)

(ii) Frame averaging in frame averaging we replaceselected watermarked frames by the average of theirprevious current and next frames The watermarkedvideo is averaged for various averaging rates and thenwe tried to extract the watermark Table 4 shows thatthe watermark can be recovered at frame averagingrates up to 50

(iii) Frame swapping the results presented in Table 5prove the robustness of our watermarking schemeagainst frame swapping because when all water-marked frames are swapped the NC is 1 and the BERis 0

Video Compression Attacks Video compression is a funda-mental attack in video watermarking that should be verifiedas video sequences are stored and transmitted in compressedformat Here we use a tool for video processing namedVirtualDub to compress the videos sequences with twodifferent Lossy compressions [31] H264 coding with a bit rateof 512kbps andMPEG4 coding with bit rates of 1500kbps and1000kbpsTheNC and BER are depicted in Table 6 and we seethat the algorithm can resist to video compression attacks

Results with the StirMark Benchmark We also evaluated theproposed method with StirMark 31 which is a well-knownevaluation tool for watermarking robustness of watermarkedvideo frames under image processing attacks [32 33] Table 7shows the evaluation of the proposed method with theStirMark benchmark under JPEG compression median filterwith JPEG compression Gaussian filter with JPEG compres-sion sharpening with JPEG compression and removal ofrow and clone with JPEG compression As can be observedthe proposed watermarking technique can resist to all thesenamed attacks

43 Comparison with Some Previously Reported AlgorithmsIn order to evaluate the performance of our algorithm wecompared the results of the proposed video watermarkingscheme to the results of related video watermarking schemesgiven in [6 16 18 20] which we introduced and discussedin the Introduction Figure 8 displays these results of com-parison under frame dropping frame averaging and JPEG

10 Security and Communication Networks

Table2Ex

tractedwatermark(w

ithNCandBE

Rvalues)u

nder

imagep

rocessingattack

Attacks

Foreman

Akiyo

Football

Noattack

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

Gaussianno

ise

NC=090

08BE

R=004

04NC=066

74BE

R=01414

NC=1

BER=0

SaltampPepp

erno

ise

NC=1

BER=0

NC=09238

BER=0303

NC=1

BER=0

Specklen

oise

NC=09479

BER=00202

NC=07870

BER=009

09NC=1

BER=0

Medianfilter

NC=09732

BER=0101

NC=1

BER=0

NC=1

BER=0

Gaussianlowpassfilter

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

Security and Communication Networks 11

Table3Ex

tractedwatermark(w

ithNCandBE

R)un

derframed

ropp

ingattack

Fram

edropp

ingrate

Foreman

Akiyo

Football

50

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

80

NC=08485

BER=006

06NC=07822

BER=0101

NC=07156

BER=01313

12 Security and Communication Networks

Table4Ex

tractedwatermark(w

ithNCandBE

R)un

derframea

veraging

attack

Fram

eaveraging

rate

Foreman

Akiyo

Football

25

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

50

NC=06819

BER=0111

1NC=0800

BER=009

09NC=1

BER=0

Security and Communication Networks 13

Table5Ex

tractedwatermark(w

ithNCandBE

R)un

derframes

wapping

attack

Fram

eSwapping

rate

Foreman

Akiyo

Football

50

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

100

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

14 Security and Communication Networks

Table6Ex

tractedwatermark(w

ithNCandBE

R)un

derv

ideo

compressio

nattack

Foreman

Akiyo

Football

H264compressio

n(512kbps)

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

MPE

G4compressio

n(1500kbp

s)NC=1

BER=0

NC=08708

BER=00505

NC=1

BER=0

MPE

G4compressio

n(100

0kbp

s)NC=1

BER=0

NC=0848

BER=006

06NC=09238

BER=00303

Security and Communication Networks 15

100

90

80

70

60

50

4010 20 30 40 50 60 70 80 90 100

NC

()

JPEG quality factor

10 20 30 40 50 60 70 80 90 100

JPEG quality factor

ForemanAkiyoFootball

ForemanAkiyoFootball

30

25

20

15

10

5

0

BER

()

Figure 7 NC and BER values of the extracted watermarks under the JPEG compression attack

Table 7 Extracted watermark (with NC and BER) under image processing attacks with StirMark benchmark

Attacks NC BERJPEG 10 05677 01515JPEG 30 09479 00202JPEG 60 08485 00606JPEG 90 10000 000003x3 median filter+ JPEG 09732 00101Gaussian filter +JPEG 07681 01010Sharpening +JPEG 05127 02626removed one row one colon +JPEG 05523 02020

compression attacks It is clear that the proposed schemeachieves a better robustness

44 Complexity of the Proposed Algorithm The proposedwatermarking schema has a low complexity because of thefollowing

(i) The algorithm is done in the MR-SVD domain whichhas an overall complexity O(n) where n indicates thesignal length [23]

(ii) To extract fast motion frames we use a simple thresh-old decision method based on motion energy

(iii) The embedding and extraction are not applied to allthe frames of the video but just to fast motion frames

In fact the execution time of the proposed scheme required inwatermark embedding and extraction processes the followingresults watermark embedding 036 s per frame watermarkextraction 019 s per frame The simulation is conducted on

Intel(R) Core i5 180 GHz processor with 4 GB RAMusingMATLAB verR2013b

5 Conclusion

In this paper we proposed a novel blind video watermarkingscheme in fast motion frames using the SVD the MR-SVD the QIM and the Logistic Map encryption In theproposed technique we solved the problem of embedding thewatermark in all frames by choosing only the frames that havebig motion energy to be the host frames suitable to the HSVUsing this approach we have limited the number of frames tobe processed and also assured a higher imperceptibility of thewatermark In addition the combination of the characteris-tics of the SVD MR-SVD QIM and Logistic Map Encryp-tion makes our scheme secure and robust to a variety ofattacks The experimental results confirm that the proposedwatermarking scheme has good imperceptibility with PSNRgreater than 40 dBTheproposed scheme is robust not only toimage processing attacks but also to frame synchronization

16 Security and Communication Networks

100

95

90

85

80

75

NC

()

NC

()

100

95

90

85

80

75

NC

()

10 15 20 25 30 35 40

frame averaging rate

proposed method[20][16][6]

proposed method

proposed method

[20]

[20]

[16][6]

[6]

[18]

10 20 30 40 50 60 70

frame dropping rate

70

65

100

95

90

85

80

75

70

65

60

JPEG Quality factor20 30 40 50 60 70 80 90

Figure 8 Comparison of experimental results between the proposed technique and other schemes under frame averaging attack framedropping attack and JPEG compression attack

and video compression attacks The comparison results withother algorithms related to video watermarking indicate thesuperiority of our scheme

Data Availability

The data used to support the findings of this study are avail-able from the corresponding author upon request

Conflicts of Interest

The authors declare no conflicts of interest

References

[1] J C IngemarM LMiller A B Jeffrey J Fridrich andT KalkerDigital Watermarking and Steganography Morgan KaufmannPublishers Inc 2008

[2] S Katzenbeisser and F Petitcolas ldquoInformation Hiding Tech-niques for Steganography and Digital Watermarkingrdquo ArtechHouse 2000

[3] H Tao L Chongmin J M Zain and A N Abdalla ldquoRobustimagewatermarking theories and techniques A reviewrdquo Journalof Applied Research and Technology vol 12 no 1 pp 122ndash1382014

Security and Communication Networks 17

[4] M Kutter F Jordan and F Bossen ldquoDigital watermarking ofcolor images using amplitude modulationrdquo Journal of ElectronicImaging vol 7 no 2 pp 326ndash332 1998

[5] A Karmakar A Phadikar B S Phadikar and G K Maity ldquoAblind video watermarking scheme resistant to rotation and col-lusion attacksrdquo Journal of King Saud University - Computer andInformation Sciences vol 28 no 2 pp 199ndash210 2016

[6] T R Singh K M Singh and S Roy ldquoVideo watermarkingscheme based on visual cryptography and scene change detec-tionrdquo AEU - International Journal of Electronics and Communi-cations vol 67 no 8 pp 645ndash651 2013

[7] W Kong B Yang D Wu and X Niu ldquoSVD Based Blind VideoWatermarking Algorithmrdquo in Proceedings of the First Interna-tional Conference on Innovative Computing Information andControl - Volume I (ICICICrsquo06) pp 265ndash268 Beijing ChinaSeptember 2006

[8] R Liu and T Tan ldquoAn SVD-based watermarking scheme forprotecting rightful ownershiprdquo IEEE Transactions on Multime-dia vol 4 no 1 pp 121ndash128 2002

[9] O S Faragallah ldquoEfficient video watermarking based on sin-gular value decomposition in the discrete wavelet transformdomainrdquo AEU - International Journal of Electronics and Com-munications vol 67 no 3 pp 189ndash196 2013

[10] K Niu X Yang and L Xiang ldquoHybrid quasi-3D DWTDCTand SVD video watermarkingrdquo in Proceedings of the 2010 IEEEInternational Conference on Software Engineering and ServiceSciences ICSESS 2010 pp 588ndash591 China July 2010

[11] N I Yassin NM Salem andM I El Adawy ldquoQIM blind videowatermarking scheme based on Wavelet transform and prin-cipal component analysisrdquo Alexandria Engineering Journal vol53 no 4 pp 833ndash842 2014

[12] D K Thind and S Jindal ldquoA semi blind DWT-SVD videowatermarkingrdquo inProceedings of the International Conference onInformation and Communication Technologies ICICT 2014 pp1661ndash1667 India December 2014

[13] N Leelavathy E V Prasad and S Kumar ldquoVideo Watermark-ing Techniques A Reviewrdquo International Journal of ComputerApplications vol 104 no 7 pp 24ndash30 2014

[14] F Hartung and M Kutter ldquoMultimedia watermarking techni-quesrdquo Proceedings of the IEEE vol 87 no 7 pp 1079ndash1107 1999

[15] P Rasti S Samiei M Agoyi S Escalera and G AnbarjafarildquoRobust non-blind color video watermarking using QR decom-position and entropy analysisrdquo Journal of Visual Communica-tion and Image Representation vol 38 pp 838ndash847 2016

[16] S M Youssef A A ElFarag and N M Ghatwary ldquoAdaptivevideo watermarking integrating a fuzzy wavelet-based humanvisual system perceptual modelrdquo Multimedia Tools and Appli-cations vol 73 no 3 pp 1545ndash1573 2014

[17] T Tabassum and S M Islam ldquoA digital video watermarkingtechnique based on identical frame extraction in 3-LevelDWTrdquoin Proceedings of the 2012 15th International Conference onComputer and Information Technology (ICCIT) pp 101ndash106Chittagong Bangladesh December 2012

[18] L Agilandeeswari and K Ganesan ldquoA robust color video water-marking scheme based on hybrid embedding techniquesrdquoMul-timedia Tools and Applications vol 75 no 14 pp 8745ndash87802016

[19] X Jiang Q Liu and Q Wu ldquoA new video watermarking algo-rithm based on shot segmentation and block classificationrdquoMultimedia Tools and Applications vol 62 no 3 pp 545ndash5602013

[20] K Chetan and K Raghavendra ldquoDWT Based Blind DigitalVideo Watermarking Scheme for Video Authenticationrdquo Inter-national Journal of Computer Applications vol 4 no 10 pp 19ndash26 2010

[21] Y JiaW Lin and A A Kassim ldquoEstimating just-noticeable dis-tortion for videordquo IEEE Transactions on Circuits and Systems forVideo Technology vol 16 no 7 pp 820ndash829 2006

[22] H C Andrews and C L Patterson ldquoSingular value decomposi-tions and digital image processingrdquo IEEE Transactions on SignalProcessing vol 24 no 1 pp 26ndash53 1976

[23] R Kakarala and P O Ogunbona ldquoSignal analysis using a multi-resolution form of the singular value decompositionrdquo IEEETransactions on Image Processing vol 10 no 5 pp 723ndash7352001

[24] B Chen and G WWornell ldquoQuantization index modulation aclass of provably goodmethods for digital watermarking and in-formation embeddingrdquo IEEE Transactions on Information The-ory vol 47 no 4 pp 1423ndash1443 2001

[25] M Yaghoobi ldquoA New Approach for Image Encryption UsingChaotic Logistic Maprdquo in Proceedings of the 2008 InternationalConference on Advanced Computer Theory and Engineering(ICACTE) pp 585ndash590 Phuket Thailand December 2008

[26] R Dugad K Ratakonda and N Ahuja ldquoRobust video shotchange detectionrdquo in Proceedings of the 1998 IEEE SecondWork-shop on Multimedia Signal Processing pp 376ndash381 RedondoBeach CA USA 1998

[27] A Anees I Hussain A Algarni and M Aslam ldquoA RobustWatermarking Scheme for Online Multimedia Copyright Pro-tection Using New Chaotic Maprdquo Security and CommunicationNetworks vol 2018 2018

[28] R Thanki V Dwivedi and K Borisagar ldquoA hybrid watermark-ing scheme with CS theory for security of multimedia datardquoJournal of King Saud University - Computer and InformationSciences 2017

[29] httpmediaxiphorgvideoderf Accessed on 06 July 2018[30] B G Haskell and A N Netravali Digital Pictures Representa-

tion Compression and Standards Perseus Publishing 1997[31] httpwwwvirtualduborg Accessed on 06 July 2018[32] F A Petitcolas R J Anderson and M G Kuhn ldquoAttacks on

CopyrightMarking Systemsrdquo in Information Hiding vol 1525 ofLecture Notes in Computer Science pp 218ndash238 Springer BerlinHeidelberg Berlin Heidelberg 1998

[33] httpwwwpetitcolasnetwatermarkingstirmark Accessed on19 September 2018

Security and Communication Networks 7

(5) TheSVD is applied to each block of the approximationcomponent01015840119896119899 = 119880119896119899 times 1198781015840119896119899times119881119879119896119899 (16)

(6) The embedded watermark is extracted by the follow-ing rule

119875 = 1 1198781015840 (1 1) minus 119876lfloor1198781015840 (1 1)119876 rfloor ge 11987620 1198781015840 (1 1) minus 119876lfloor1198781015840 (1 1)119876 rfloor lt 1198762 (17)

where lfloor∙rfloor is the floor function(7) Decryption with the same chaotic sequence is per-

formed to get the hidden binary watermark W

(8) Since a video clip contains several fast motion framesin which the same watermark is embedded we calcu-late the final recovered watermark by averaging thewatermarks extracted from these different frames

4 Results and Discussion

Theproposed algorithm is implemented using MATLABWeused three CIF (352times288) standard sequences shown in Fig-ure 4 (Foreman 300 frames Akiyo 300 frames and Football260 frames) The test videos are in RGB uncompressed aviformat with a frame rate of 30 fps [29] The watermark is abinary image Its resolution is 9 times 11 The block size in theproposed algorithm can be set to any desired value Howevera small block size leads to block effects problem whereas alarge block size reduces the total number of the watermarkbits We carried out various experiments and found that theblock size of 16x16 allows the embedding of an acceptablenumber of watermark bits without causing noticeable blockeffect

As shown inFigure 5 small values of the quantization stepyield good transparency at the expense of poor robustnessand vice versa From this figure it can be observed that thevalue of the quantization step Q = 110 gives a good com-promise between robustness and imperceptibility

41 Imperceptibility Tests The imperceptibility of the water-mark is estimated by measuring the PSNR (Peak Signal toNoise Ratio) and Mean Structure Similarity Index Measure(MSSIM) which are calculated using the luminance space Yof the original and watermarked frames [30] High PSNRvalues of the watermarked video frames indicate betterimperceptibility It is worth noting that the SSIM provides aperceptual distortion in range of [0 1] When both framesare numerically the same this value is equal to 1 The PSNRis calculated as follows

119875119878119873119877 = 10 log10 ( 2562119872119878119864) (18)

Table 1 The PSNR and MSSIM of the watermarked videos

Video Foreman Akiyo FootballAverage PSNR 4020 4045 4026Average MSSIM 09981 09988 09977

where the Mean Square Error (MSE) between the host lumi-nance Y and the watermarked luminance Y1015840 is defined as

119872119878119864 = 1119872 times119873119872minus1sum119894=0 119873minus1sum119895=0 1003816100381610038161003816119884 (119894 119895) minus 1198841015840 (119894 119895)100381610038161003816100381610038162 (19)

with M and N respectively being the height and width of thevideo frame

The MSSIM is defined as follows

119872119878119878119868119872(119884 1198841015840) = 1119872 119872sum119895=1 119878119878119868119872(119884119869 1198841015840119895) (20)

where YJ and Y1015840j are the image contents at the j local windowand M is the number of local windows of the image

119878119878119868119872(119884 1198841015840) = [119897 (119884 1198841015840)]120572 sdot [119888 (119884 1198841015840)]120573 sdot 119904 [119884 1198841015840)]120574 (21)

where l c and s are the luminance comparison the contrastcomparison and the structure comparison functions respec-tively With 120572 120573 and 120574 are parameters used to adjust therelative importance of the three components

As shown inTable 1 the average PSNRvalues of thewater-marked videos are higher than 40 dB and the correspondingMSSIMvalues are very close to 1This indicates the invisibilityof the watermark which means that the watermarked videosappear visually identical to the original ones as shown inFigure 6

42 Robustness Tests The robustness for any watermarkingsystem is a very important requirement To verify it weapply to the watermarked video various types of attacks andwe use the Normalized Coefficient (NC) and the Bit ErrorRate (BER) to compare the similarities between the originalwatermark W and the extracted watermark W The NC andBER are respectively calculated as

119873119862(119882 ) = sumsum 119882(119894 119895) (119894 119895)1003816100381610038161003816119882 (119894 119895)10038161003816100381610038162 (22)

119861119864119877 = 1119875 119901sum119895=1

10038161003816100381610038161003816 (119895) minus 119882 (119895)10038161003816100381610038161003816 (23)

where W W and P are respectively the original watermarkthe extracted watermark and the size of the watermark

The correlation between W and W is very high when NCis close to 1

Generally in video watermarking attacks are divided intothree categories image processing attacks frame synchro-nization attacks and video compression attacks

8 Security and Communication Networks

(a) (b)

(c)

Figure 4 Video sequences used for testing (a) Foreman (b) Akiyo and (c) Football

1

09965

0997

09975

0998

09985

0999

09995

MSS

IM

60 80 100 120 14040Q-step

(a)

60 80 100 120 14040Q-step

055

06

065

07

075

08

085

09

095

1

NC

(b)

Figure 5 (a) MSSIM of watermarked video for different quantization step Q and (b) NC of extracted watermark with JPEG compressionattack for different quantization step Q

Image Processing Attacks Considering a video as a sequenceof images the attacks applied to images can then be applied tothe video sequences The common image processing attacksare as follows

(i) Adding a noise three kinds of noises are added tothe watermarked video Gaussian noise salt amp peppernoise and speckle noise with density of 1 It canbe seen from Table 2 that the watermark is always

Security and Communication Networks 9

(a) (b)

(c)

Figure 6 Original and watermarked frames (a) frame 228 of Foreman (b) frame 16 of Akiyo and (c) frame 128 of Football

detectablewithNCandBER values respectively closeto 1 and 0 especially for Foreman and Football videosequences

(ii) Filtering 3x3median filter and 5x5 Gaussian filter areapplied separately to the watermarked videos and wecan see from Table 2 that the proposed method isrobust against median and Gaussian filtering

(iii) JPEG compression the watermarked video is com-pressed with different quality factors ranging from 10to 100 Figure 7 shows the results for the JPEG com-pression for instance if the watermarked video iscompressed with a quality factor of 40 the obtainedNC is greater than 95 and the BER value is lowerthan 02 This confirms the robustness of the pro-posed scheme to the JPEG compression attack

Frame Synchronization Attacks Because contents in the con-secutive frames of a video are almost identical it makesthe video sequences susceptible to temporal synchronizationattacks such as frame averaging frame dropping and frameswapping

(i) Frame dropping in frame dropping selected water-marked frames are replaced by their correspondingoriginal frames Table 3 shows the average NC andBER values given at different frame dropping ratesOur scheme achieves strong robustness against framedropping even for the case of high rates (ie 80)

(ii) Frame averaging in frame averaging we replaceselected watermarked frames by the average of theirprevious current and next frames The watermarkedvideo is averaged for various averaging rates and thenwe tried to extract the watermark Table 4 shows thatthe watermark can be recovered at frame averagingrates up to 50

(iii) Frame swapping the results presented in Table 5prove the robustness of our watermarking schemeagainst frame swapping because when all water-marked frames are swapped the NC is 1 and the BERis 0

Video Compression Attacks Video compression is a funda-mental attack in video watermarking that should be verifiedas video sequences are stored and transmitted in compressedformat Here we use a tool for video processing namedVirtualDub to compress the videos sequences with twodifferent Lossy compressions [31] H264 coding with a bit rateof 512kbps andMPEG4 coding with bit rates of 1500kbps and1000kbpsTheNC and BER are depicted in Table 6 and we seethat the algorithm can resist to video compression attacks

Results with the StirMark Benchmark We also evaluated theproposed method with StirMark 31 which is a well-knownevaluation tool for watermarking robustness of watermarkedvideo frames under image processing attacks [32 33] Table 7shows the evaluation of the proposed method with theStirMark benchmark under JPEG compression median filterwith JPEG compression Gaussian filter with JPEG compres-sion sharpening with JPEG compression and removal ofrow and clone with JPEG compression As can be observedthe proposed watermarking technique can resist to all thesenamed attacks

43 Comparison with Some Previously Reported AlgorithmsIn order to evaluate the performance of our algorithm wecompared the results of the proposed video watermarkingscheme to the results of related video watermarking schemesgiven in [6 16 18 20] which we introduced and discussedin the Introduction Figure 8 displays these results of com-parison under frame dropping frame averaging and JPEG

10 Security and Communication Networks

Table2Ex

tractedwatermark(w

ithNCandBE

Rvalues)u

nder

imagep

rocessingattack

Attacks

Foreman

Akiyo

Football

Noattack

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

Gaussianno

ise

NC=090

08BE

R=004

04NC=066

74BE

R=01414

NC=1

BER=0

SaltampPepp

erno

ise

NC=1

BER=0

NC=09238

BER=0303

NC=1

BER=0

Specklen

oise

NC=09479

BER=00202

NC=07870

BER=009

09NC=1

BER=0

Medianfilter

NC=09732

BER=0101

NC=1

BER=0

NC=1

BER=0

Gaussianlowpassfilter

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

Security and Communication Networks 11

Table3Ex

tractedwatermark(w

ithNCandBE

R)un

derframed

ropp

ingattack

Fram

edropp

ingrate

Foreman

Akiyo

Football

50

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

80

NC=08485

BER=006

06NC=07822

BER=0101

NC=07156

BER=01313

12 Security and Communication Networks

Table4Ex

tractedwatermark(w

ithNCandBE

R)un

derframea

veraging

attack

Fram

eaveraging

rate

Foreman

Akiyo

Football

25

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

50

NC=06819

BER=0111

1NC=0800

BER=009

09NC=1

BER=0

Security and Communication Networks 13

Table5Ex

tractedwatermark(w

ithNCandBE

R)un

derframes

wapping

attack

Fram

eSwapping

rate

Foreman

Akiyo

Football

50

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

100

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

14 Security and Communication Networks

Table6Ex

tractedwatermark(w

ithNCandBE

R)un

derv

ideo

compressio

nattack

Foreman

Akiyo

Football

H264compressio

n(512kbps)

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

MPE

G4compressio

n(1500kbp

s)NC=1

BER=0

NC=08708

BER=00505

NC=1

BER=0

MPE

G4compressio

n(100

0kbp

s)NC=1

BER=0

NC=0848

BER=006

06NC=09238

BER=00303

Security and Communication Networks 15

100

90

80

70

60

50

4010 20 30 40 50 60 70 80 90 100

NC

()

JPEG quality factor

10 20 30 40 50 60 70 80 90 100

JPEG quality factor

ForemanAkiyoFootball

ForemanAkiyoFootball

30

25

20

15

10

5

0

BER

()

Figure 7 NC and BER values of the extracted watermarks under the JPEG compression attack

Table 7 Extracted watermark (with NC and BER) under image processing attacks with StirMark benchmark

Attacks NC BERJPEG 10 05677 01515JPEG 30 09479 00202JPEG 60 08485 00606JPEG 90 10000 000003x3 median filter+ JPEG 09732 00101Gaussian filter +JPEG 07681 01010Sharpening +JPEG 05127 02626removed one row one colon +JPEG 05523 02020

compression attacks It is clear that the proposed schemeachieves a better robustness

44 Complexity of the Proposed Algorithm The proposedwatermarking schema has a low complexity because of thefollowing

(i) The algorithm is done in the MR-SVD domain whichhas an overall complexity O(n) where n indicates thesignal length [23]

(ii) To extract fast motion frames we use a simple thresh-old decision method based on motion energy

(iii) The embedding and extraction are not applied to allthe frames of the video but just to fast motion frames

In fact the execution time of the proposed scheme required inwatermark embedding and extraction processes the followingresults watermark embedding 036 s per frame watermarkextraction 019 s per frame The simulation is conducted on

Intel(R) Core i5 180 GHz processor with 4 GB RAMusingMATLAB verR2013b

5 Conclusion

In this paper we proposed a novel blind video watermarkingscheme in fast motion frames using the SVD the MR-SVD the QIM and the Logistic Map encryption In theproposed technique we solved the problem of embedding thewatermark in all frames by choosing only the frames that havebig motion energy to be the host frames suitable to the HSVUsing this approach we have limited the number of frames tobe processed and also assured a higher imperceptibility of thewatermark In addition the combination of the characteris-tics of the SVD MR-SVD QIM and Logistic Map Encryp-tion makes our scheme secure and robust to a variety ofattacks The experimental results confirm that the proposedwatermarking scheme has good imperceptibility with PSNRgreater than 40 dBTheproposed scheme is robust not only toimage processing attacks but also to frame synchronization

16 Security and Communication Networks

100

95

90

85

80

75

NC

()

NC

()

100

95

90

85

80

75

NC

()

10 15 20 25 30 35 40

frame averaging rate

proposed method[20][16][6]

proposed method

proposed method

[20]

[20]

[16][6]

[6]

[18]

10 20 30 40 50 60 70

frame dropping rate

70

65

100

95

90

85

80

75

70

65

60

JPEG Quality factor20 30 40 50 60 70 80 90

Figure 8 Comparison of experimental results between the proposed technique and other schemes under frame averaging attack framedropping attack and JPEG compression attack

and video compression attacks The comparison results withother algorithms related to video watermarking indicate thesuperiority of our scheme

Data Availability

The data used to support the findings of this study are avail-able from the corresponding author upon request

Conflicts of Interest

The authors declare no conflicts of interest

References

[1] J C IngemarM LMiller A B Jeffrey J Fridrich andT KalkerDigital Watermarking and Steganography Morgan KaufmannPublishers Inc 2008

[2] S Katzenbeisser and F Petitcolas ldquoInformation Hiding Tech-niques for Steganography and Digital Watermarkingrdquo ArtechHouse 2000

[3] H Tao L Chongmin J M Zain and A N Abdalla ldquoRobustimagewatermarking theories and techniques A reviewrdquo Journalof Applied Research and Technology vol 12 no 1 pp 122ndash1382014

Security and Communication Networks 17

[4] M Kutter F Jordan and F Bossen ldquoDigital watermarking ofcolor images using amplitude modulationrdquo Journal of ElectronicImaging vol 7 no 2 pp 326ndash332 1998

[5] A Karmakar A Phadikar B S Phadikar and G K Maity ldquoAblind video watermarking scheme resistant to rotation and col-lusion attacksrdquo Journal of King Saud University - Computer andInformation Sciences vol 28 no 2 pp 199ndash210 2016

[6] T R Singh K M Singh and S Roy ldquoVideo watermarkingscheme based on visual cryptography and scene change detec-tionrdquo AEU - International Journal of Electronics and Communi-cations vol 67 no 8 pp 645ndash651 2013

[7] W Kong B Yang D Wu and X Niu ldquoSVD Based Blind VideoWatermarking Algorithmrdquo in Proceedings of the First Interna-tional Conference on Innovative Computing Information andControl - Volume I (ICICICrsquo06) pp 265ndash268 Beijing ChinaSeptember 2006

[8] R Liu and T Tan ldquoAn SVD-based watermarking scheme forprotecting rightful ownershiprdquo IEEE Transactions on Multime-dia vol 4 no 1 pp 121ndash128 2002

[9] O S Faragallah ldquoEfficient video watermarking based on sin-gular value decomposition in the discrete wavelet transformdomainrdquo AEU - International Journal of Electronics and Com-munications vol 67 no 3 pp 189ndash196 2013

[10] K Niu X Yang and L Xiang ldquoHybrid quasi-3D DWTDCTand SVD video watermarkingrdquo in Proceedings of the 2010 IEEEInternational Conference on Software Engineering and ServiceSciences ICSESS 2010 pp 588ndash591 China July 2010

[11] N I Yassin NM Salem andM I El Adawy ldquoQIM blind videowatermarking scheme based on Wavelet transform and prin-cipal component analysisrdquo Alexandria Engineering Journal vol53 no 4 pp 833ndash842 2014

[12] D K Thind and S Jindal ldquoA semi blind DWT-SVD videowatermarkingrdquo inProceedings of the International Conference onInformation and Communication Technologies ICICT 2014 pp1661ndash1667 India December 2014

[13] N Leelavathy E V Prasad and S Kumar ldquoVideo Watermark-ing Techniques A Reviewrdquo International Journal of ComputerApplications vol 104 no 7 pp 24ndash30 2014

[14] F Hartung and M Kutter ldquoMultimedia watermarking techni-quesrdquo Proceedings of the IEEE vol 87 no 7 pp 1079ndash1107 1999

[15] P Rasti S Samiei M Agoyi S Escalera and G AnbarjafarildquoRobust non-blind color video watermarking using QR decom-position and entropy analysisrdquo Journal of Visual Communica-tion and Image Representation vol 38 pp 838ndash847 2016

[16] S M Youssef A A ElFarag and N M Ghatwary ldquoAdaptivevideo watermarking integrating a fuzzy wavelet-based humanvisual system perceptual modelrdquo Multimedia Tools and Appli-cations vol 73 no 3 pp 1545ndash1573 2014

[17] T Tabassum and S M Islam ldquoA digital video watermarkingtechnique based on identical frame extraction in 3-LevelDWTrdquoin Proceedings of the 2012 15th International Conference onComputer and Information Technology (ICCIT) pp 101ndash106Chittagong Bangladesh December 2012

[18] L Agilandeeswari and K Ganesan ldquoA robust color video water-marking scheme based on hybrid embedding techniquesrdquoMul-timedia Tools and Applications vol 75 no 14 pp 8745ndash87802016

[19] X Jiang Q Liu and Q Wu ldquoA new video watermarking algo-rithm based on shot segmentation and block classificationrdquoMultimedia Tools and Applications vol 62 no 3 pp 545ndash5602013

[20] K Chetan and K Raghavendra ldquoDWT Based Blind DigitalVideo Watermarking Scheme for Video Authenticationrdquo Inter-national Journal of Computer Applications vol 4 no 10 pp 19ndash26 2010

[21] Y JiaW Lin and A A Kassim ldquoEstimating just-noticeable dis-tortion for videordquo IEEE Transactions on Circuits and Systems forVideo Technology vol 16 no 7 pp 820ndash829 2006

[22] H C Andrews and C L Patterson ldquoSingular value decomposi-tions and digital image processingrdquo IEEE Transactions on SignalProcessing vol 24 no 1 pp 26ndash53 1976

[23] R Kakarala and P O Ogunbona ldquoSignal analysis using a multi-resolution form of the singular value decompositionrdquo IEEETransactions on Image Processing vol 10 no 5 pp 723ndash7352001

[24] B Chen and G WWornell ldquoQuantization index modulation aclass of provably goodmethods for digital watermarking and in-formation embeddingrdquo IEEE Transactions on Information The-ory vol 47 no 4 pp 1423ndash1443 2001

[25] M Yaghoobi ldquoA New Approach for Image Encryption UsingChaotic Logistic Maprdquo in Proceedings of the 2008 InternationalConference on Advanced Computer Theory and Engineering(ICACTE) pp 585ndash590 Phuket Thailand December 2008

[26] R Dugad K Ratakonda and N Ahuja ldquoRobust video shotchange detectionrdquo in Proceedings of the 1998 IEEE SecondWork-shop on Multimedia Signal Processing pp 376ndash381 RedondoBeach CA USA 1998

[27] A Anees I Hussain A Algarni and M Aslam ldquoA RobustWatermarking Scheme for Online Multimedia Copyright Pro-tection Using New Chaotic Maprdquo Security and CommunicationNetworks vol 2018 2018

[28] R Thanki V Dwivedi and K Borisagar ldquoA hybrid watermark-ing scheme with CS theory for security of multimedia datardquoJournal of King Saud University - Computer and InformationSciences 2017

[29] httpmediaxiphorgvideoderf Accessed on 06 July 2018[30] B G Haskell and A N Netravali Digital Pictures Representa-

tion Compression and Standards Perseus Publishing 1997[31] httpwwwvirtualduborg Accessed on 06 July 2018[32] F A Petitcolas R J Anderson and M G Kuhn ldquoAttacks on

CopyrightMarking Systemsrdquo in Information Hiding vol 1525 ofLecture Notes in Computer Science pp 218ndash238 Springer BerlinHeidelberg Berlin Heidelberg 1998

[33] httpwwwpetitcolasnetwatermarkingstirmark Accessed on19 September 2018

8 Security and Communication Networks

(a) (b)

(c)

Figure 4 Video sequences used for testing (a) Foreman (b) Akiyo and (c) Football

1

09965

0997

09975

0998

09985

0999

09995

MSS

IM

60 80 100 120 14040Q-step

(a)

60 80 100 120 14040Q-step

055

06

065

07

075

08

085

09

095

1

NC

(b)

Figure 5 (a) MSSIM of watermarked video for different quantization step Q and (b) NC of extracted watermark with JPEG compressionattack for different quantization step Q

Image Processing Attacks Considering a video as a sequenceof images the attacks applied to images can then be applied tothe video sequences The common image processing attacksare as follows

(i) Adding a noise three kinds of noises are added tothe watermarked video Gaussian noise salt amp peppernoise and speckle noise with density of 1 It canbe seen from Table 2 that the watermark is always

Security and Communication Networks 9

(a) (b)

(c)

Figure 6 Original and watermarked frames (a) frame 228 of Foreman (b) frame 16 of Akiyo and (c) frame 128 of Football

detectablewithNCandBER values respectively closeto 1 and 0 especially for Foreman and Football videosequences

(ii) Filtering 3x3median filter and 5x5 Gaussian filter areapplied separately to the watermarked videos and wecan see from Table 2 that the proposed method isrobust against median and Gaussian filtering

(iii) JPEG compression the watermarked video is com-pressed with different quality factors ranging from 10to 100 Figure 7 shows the results for the JPEG com-pression for instance if the watermarked video iscompressed with a quality factor of 40 the obtainedNC is greater than 95 and the BER value is lowerthan 02 This confirms the robustness of the pro-posed scheme to the JPEG compression attack

Frame Synchronization Attacks Because contents in the con-secutive frames of a video are almost identical it makesthe video sequences susceptible to temporal synchronizationattacks such as frame averaging frame dropping and frameswapping

(i) Frame dropping in frame dropping selected water-marked frames are replaced by their correspondingoriginal frames Table 3 shows the average NC andBER values given at different frame dropping ratesOur scheme achieves strong robustness against framedropping even for the case of high rates (ie 80)

(ii) Frame averaging in frame averaging we replaceselected watermarked frames by the average of theirprevious current and next frames The watermarkedvideo is averaged for various averaging rates and thenwe tried to extract the watermark Table 4 shows thatthe watermark can be recovered at frame averagingrates up to 50

(iii) Frame swapping the results presented in Table 5prove the robustness of our watermarking schemeagainst frame swapping because when all water-marked frames are swapped the NC is 1 and the BERis 0

Video Compression Attacks Video compression is a funda-mental attack in video watermarking that should be verifiedas video sequences are stored and transmitted in compressedformat Here we use a tool for video processing namedVirtualDub to compress the videos sequences with twodifferent Lossy compressions [31] H264 coding with a bit rateof 512kbps andMPEG4 coding with bit rates of 1500kbps and1000kbpsTheNC and BER are depicted in Table 6 and we seethat the algorithm can resist to video compression attacks

Results with the StirMark Benchmark We also evaluated theproposed method with StirMark 31 which is a well-knownevaluation tool for watermarking robustness of watermarkedvideo frames under image processing attacks [32 33] Table 7shows the evaluation of the proposed method with theStirMark benchmark under JPEG compression median filterwith JPEG compression Gaussian filter with JPEG compres-sion sharpening with JPEG compression and removal ofrow and clone with JPEG compression As can be observedthe proposed watermarking technique can resist to all thesenamed attacks

43 Comparison with Some Previously Reported AlgorithmsIn order to evaluate the performance of our algorithm wecompared the results of the proposed video watermarkingscheme to the results of related video watermarking schemesgiven in [6 16 18 20] which we introduced and discussedin the Introduction Figure 8 displays these results of com-parison under frame dropping frame averaging and JPEG

10 Security and Communication Networks

Table2Ex

tractedwatermark(w

ithNCandBE

Rvalues)u

nder

imagep

rocessingattack

Attacks

Foreman

Akiyo

Football

Noattack

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

Gaussianno

ise

NC=090

08BE

R=004

04NC=066

74BE

R=01414

NC=1

BER=0

SaltampPepp

erno

ise

NC=1

BER=0

NC=09238

BER=0303

NC=1

BER=0

Specklen

oise

NC=09479

BER=00202

NC=07870

BER=009

09NC=1

BER=0

Medianfilter

NC=09732

BER=0101

NC=1

BER=0

NC=1

BER=0

Gaussianlowpassfilter

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

Security and Communication Networks 11

Table3Ex

tractedwatermark(w

ithNCandBE

R)un

derframed

ropp

ingattack

Fram

edropp

ingrate

Foreman

Akiyo

Football

50

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

80

NC=08485

BER=006

06NC=07822

BER=0101

NC=07156

BER=01313

12 Security and Communication Networks

Table4Ex

tractedwatermark(w

ithNCandBE

R)un

derframea

veraging

attack

Fram

eaveraging

rate

Foreman

Akiyo

Football

25

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

50

NC=06819

BER=0111

1NC=0800

BER=009

09NC=1

BER=0

Security and Communication Networks 13

Table5Ex

tractedwatermark(w

ithNCandBE

R)un

derframes

wapping

attack

Fram

eSwapping

rate

Foreman

Akiyo

Football

50

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

100

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

14 Security and Communication Networks

Table6Ex

tractedwatermark(w

ithNCandBE

R)un

derv

ideo

compressio

nattack

Foreman

Akiyo

Football

H264compressio

n(512kbps)

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

MPE

G4compressio

n(1500kbp

s)NC=1

BER=0

NC=08708

BER=00505

NC=1

BER=0

MPE

G4compressio

n(100

0kbp

s)NC=1

BER=0

NC=0848

BER=006

06NC=09238

BER=00303

Security and Communication Networks 15

100

90

80

70

60

50

4010 20 30 40 50 60 70 80 90 100

NC

()

JPEG quality factor

10 20 30 40 50 60 70 80 90 100

JPEG quality factor

ForemanAkiyoFootball

ForemanAkiyoFootball

30

25

20

15

10

5

0

BER

()

Figure 7 NC and BER values of the extracted watermarks under the JPEG compression attack

Table 7 Extracted watermark (with NC and BER) under image processing attacks with StirMark benchmark

Attacks NC BERJPEG 10 05677 01515JPEG 30 09479 00202JPEG 60 08485 00606JPEG 90 10000 000003x3 median filter+ JPEG 09732 00101Gaussian filter +JPEG 07681 01010Sharpening +JPEG 05127 02626removed one row one colon +JPEG 05523 02020

compression attacks It is clear that the proposed schemeachieves a better robustness

44 Complexity of the Proposed Algorithm The proposedwatermarking schema has a low complexity because of thefollowing

(i) The algorithm is done in the MR-SVD domain whichhas an overall complexity O(n) where n indicates thesignal length [23]

(ii) To extract fast motion frames we use a simple thresh-old decision method based on motion energy

(iii) The embedding and extraction are not applied to allthe frames of the video but just to fast motion frames

In fact the execution time of the proposed scheme required inwatermark embedding and extraction processes the followingresults watermark embedding 036 s per frame watermarkextraction 019 s per frame The simulation is conducted on

Intel(R) Core i5 180 GHz processor with 4 GB RAMusingMATLAB verR2013b

5 Conclusion

In this paper we proposed a novel blind video watermarkingscheme in fast motion frames using the SVD the MR-SVD the QIM and the Logistic Map encryption In theproposed technique we solved the problem of embedding thewatermark in all frames by choosing only the frames that havebig motion energy to be the host frames suitable to the HSVUsing this approach we have limited the number of frames tobe processed and also assured a higher imperceptibility of thewatermark In addition the combination of the characteris-tics of the SVD MR-SVD QIM and Logistic Map Encryp-tion makes our scheme secure and robust to a variety ofattacks The experimental results confirm that the proposedwatermarking scheme has good imperceptibility with PSNRgreater than 40 dBTheproposed scheme is robust not only toimage processing attacks but also to frame synchronization

16 Security and Communication Networks

100

95

90

85

80

75

NC

()

NC

()

100

95

90

85

80

75

NC

()

10 15 20 25 30 35 40

frame averaging rate

proposed method[20][16][6]

proposed method

proposed method

[20]

[20]

[16][6]

[6]

[18]

10 20 30 40 50 60 70

frame dropping rate

70

65

100

95

90

85

80

75

70

65

60

JPEG Quality factor20 30 40 50 60 70 80 90

Figure 8 Comparison of experimental results between the proposed technique and other schemes under frame averaging attack framedropping attack and JPEG compression attack

and video compression attacks The comparison results withother algorithms related to video watermarking indicate thesuperiority of our scheme

Data Availability

The data used to support the findings of this study are avail-able from the corresponding author upon request

Conflicts of Interest

The authors declare no conflicts of interest

References

[1] J C IngemarM LMiller A B Jeffrey J Fridrich andT KalkerDigital Watermarking and Steganography Morgan KaufmannPublishers Inc 2008

[2] S Katzenbeisser and F Petitcolas ldquoInformation Hiding Tech-niques for Steganography and Digital Watermarkingrdquo ArtechHouse 2000

[3] H Tao L Chongmin J M Zain and A N Abdalla ldquoRobustimagewatermarking theories and techniques A reviewrdquo Journalof Applied Research and Technology vol 12 no 1 pp 122ndash1382014

Security and Communication Networks 17

[4] M Kutter F Jordan and F Bossen ldquoDigital watermarking ofcolor images using amplitude modulationrdquo Journal of ElectronicImaging vol 7 no 2 pp 326ndash332 1998

[5] A Karmakar A Phadikar B S Phadikar and G K Maity ldquoAblind video watermarking scheme resistant to rotation and col-lusion attacksrdquo Journal of King Saud University - Computer andInformation Sciences vol 28 no 2 pp 199ndash210 2016

[6] T R Singh K M Singh and S Roy ldquoVideo watermarkingscheme based on visual cryptography and scene change detec-tionrdquo AEU - International Journal of Electronics and Communi-cations vol 67 no 8 pp 645ndash651 2013

[7] W Kong B Yang D Wu and X Niu ldquoSVD Based Blind VideoWatermarking Algorithmrdquo in Proceedings of the First Interna-tional Conference on Innovative Computing Information andControl - Volume I (ICICICrsquo06) pp 265ndash268 Beijing ChinaSeptember 2006

[8] R Liu and T Tan ldquoAn SVD-based watermarking scheme forprotecting rightful ownershiprdquo IEEE Transactions on Multime-dia vol 4 no 1 pp 121ndash128 2002

[9] O S Faragallah ldquoEfficient video watermarking based on sin-gular value decomposition in the discrete wavelet transformdomainrdquo AEU - International Journal of Electronics and Com-munications vol 67 no 3 pp 189ndash196 2013

[10] K Niu X Yang and L Xiang ldquoHybrid quasi-3D DWTDCTand SVD video watermarkingrdquo in Proceedings of the 2010 IEEEInternational Conference on Software Engineering and ServiceSciences ICSESS 2010 pp 588ndash591 China July 2010

[11] N I Yassin NM Salem andM I El Adawy ldquoQIM blind videowatermarking scheme based on Wavelet transform and prin-cipal component analysisrdquo Alexandria Engineering Journal vol53 no 4 pp 833ndash842 2014

[12] D K Thind and S Jindal ldquoA semi blind DWT-SVD videowatermarkingrdquo inProceedings of the International Conference onInformation and Communication Technologies ICICT 2014 pp1661ndash1667 India December 2014

[13] N Leelavathy E V Prasad and S Kumar ldquoVideo Watermark-ing Techniques A Reviewrdquo International Journal of ComputerApplications vol 104 no 7 pp 24ndash30 2014

[14] F Hartung and M Kutter ldquoMultimedia watermarking techni-quesrdquo Proceedings of the IEEE vol 87 no 7 pp 1079ndash1107 1999

[15] P Rasti S Samiei M Agoyi S Escalera and G AnbarjafarildquoRobust non-blind color video watermarking using QR decom-position and entropy analysisrdquo Journal of Visual Communica-tion and Image Representation vol 38 pp 838ndash847 2016

[16] S M Youssef A A ElFarag and N M Ghatwary ldquoAdaptivevideo watermarking integrating a fuzzy wavelet-based humanvisual system perceptual modelrdquo Multimedia Tools and Appli-cations vol 73 no 3 pp 1545ndash1573 2014

[17] T Tabassum and S M Islam ldquoA digital video watermarkingtechnique based on identical frame extraction in 3-LevelDWTrdquoin Proceedings of the 2012 15th International Conference onComputer and Information Technology (ICCIT) pp 101ndash106Chittagong Bangladesh December 2012

[18] L Agilandeeswari and K Ganesan ldquoA robust color video water-marking scheme based on hybrid embedding techniquesrdquoMul-timedia Tools and Applications vol 75 no 14 pp 8745ndash87802016

[19] X Jiang Q Liu and Q Wu ldquoA new video watermarking algo-rithm based on shot segmentation and block classificationrdquoMultimedia Tools and Applications vol 62 no 3 pp 545ndash5602013

[20] K Chetan and K Raghavendra ldquoDWT Based Blind DigitalVideo Watermarking Scheme for Video Authenticationrdquo Inter-national Journal of Computer Applications vol 4 no 10 pp 19ndash26 2010

[21] Y JiaW Lin and A A Kassim ldquoEstimating just-noticeable dis-tortion for videordquo IEEE Transactions on Circuits and Systems forVideo Technology vol 16 no 7 pp 820ndash829 2006

[22] H C Andrews and C L Patterson ldquoSingular value decomposi-tions and digital image processingrdquo IEEE Transactions on SignalProcessing vol 24 no 1 pp 26ndash53 1976

[23] R Kakarala and P O Ogunbona ldquoSignal analysis using a multi-resolution form of the singular value decompositionrdquo IEEETransactions on Image Processing vol 10 no 5 pp 723ndash7352001

[24] B Chen and G WWornell ldquoQuantization index modulation aclass of provably goodmethods for digital watermarking and in-formation embeddingrdquo IEEE Transactions on Information The-ory vol 47 no 4 pp 1423ndash1443 2001

[25] M Yaghoobi ldquoA New Approach for Image Encryption UsingChaotic Logistic Maprdquo in Proceedings of the 2008 InternationalConference on Advanced Computer Theory and Engineering(ICACTE) pp 585ndash590 Phuket Thailand December 2008

[26] R Dugad K Ratakonda and N Ahuja ldquoRobust video shotchange detectionrdquo in Proceedings of the 1998 IEEE SecondWork-shop on Multimedia Signal Processing pp 376ndash381 RedondoBeach CA USA 1998

[27] A Anees I Hussain A Algarni and M Aslam ldquoA RobustWatermarking Scheme for Online Multimedia Copyright Pro-tection Using New Chaotic Maprdquo Security and CommunicationNetworks vol 2018 2018

[28] R Thanki V Dwivedi and K Borisagar ldquoA hybrid watermark-ing scheme with CS theory for security of multimedia datardquoJournal of King Saud University - Computer and InformationSciences 2017

[29] httpmediaxiphorgvideoderf Accessed on 06 July 2018[30] B G Haskell and A N Netravali Digital Pictures Representa-

tion Compression and Standards Perseus Publishing 1997[31] httpwwwvirtualduborg Accessed on 06 July 2018[32] F A Petitcolas R J Anderson and M G Kuhn ldquoAttacks on

CopyrightMarking Systemsrdquo in Information Hiding vol 1525 ofLecture Notes in Computer Science pp 218ndash238 Springer BerlinHeidelberg Berlin Heidelberg 1998

[33] httpwwwpetitcolasnetwatermarkingstirmark Accessed on19 September 2018

Security and Communication Networks 9

(a) (b)

(c)

Figure 6 Original and watermarked frames (a) frame 228 of Foreman (b) frame 16 of Akiyo and (c) frame 128 of Football

detectablewithNCandBER values respectively closeto 1 and 0 especially for Foreman and Football videosequences

(ii) Filtering 3x3median filter and 5x5 Gaussian filter areapplied separately to the watermarked videos and wecan see from Table 2 that the proposed method isrobust against median and Gaussian filtering

(iii) JPEG compression the watermarked video is com-pressed with different quality factors ranging from 10to 100 Figure 7 shows the results for the JPEG com-pression for instance if the watermarked video iscompressed with a quality factor of 40 the obtainedNC is greater than 95 and the BER value is lowerthan 02 This confirms the robustness of the pro-posed scheme to the JPEG compression attack

Frame Synchronization Attacks Because contents in the con-secutive frames of a video are almost identical it makesthe video sequences susceptible to temporal synchronizationattacks such as frame averaging frame dropping and frameswapping

(i) Frame dropping in frame dropping selected water-marked frames are replaced by their correspondingoriginal frames Table 3 shows the average NC andBER values given at different frame dropping ratesOur scheme achieves strong robustness against framedropping even for the case of high rates (ie 80)

(ii) Frame averaging in frame averaging we replaceselected watermarked frames by the average of theirprevious current and next frames The watermarkedvideo is averaged for various averaging rates and thenwe tried to extract the watermark Table 4 shows thatthe watermark can be recovered at frame averagingrates up to 50

(iii) Frame swapping the results presented in Table 5prove the robustness of our watermarking schemeagainst frame swapping because when all water-marked frames are swapped the NC is 1 and the BERis 0

Video Compression Attacks Video compression is a funda-mental attack in video watermarking that should be verifiedas video sequences are stored and transmitted in compressedformat Here we use a tool for video processing namedVirtualDub to compress the videos sequences with twodifferent Lossy compressions [31] H264 coding with a bit rateof 512kbps andMPEG4 coding with bit rates of 1500kbps and1000kbpsTheNC and BER are depicted in Table 6 and we seethat the algorithm can resist to video compression attacks

Results with the StirMark Benchmark We also evaluated theproposed method with StirMark 31 which is a well-knownevaluation tool for watermarking robustness of watermarkedvideo frames under image processing attacks [32 33] Table 7shows the evaluation of the proposed method with theStirMark benchmark under JPEG compression median filterwith JPEG compression Gaussian filter with JPEG compres-sion sharpening with JPEG compression and removal ofrow and clone with JPEG compression As can be observedthe proposed watermarking technique can resist to all thesenamed attacks

43 Comparison with Some Previously Reported AlgorithmsIn order to evaluate the performance of our algorithm wecompared the results of the proposed video watermarkingscheme to the results of related video watermarking schemesgiven in [6 16 18 20] which we introduced and discussedin the Introduction Figure 8 displays these results of com-parison under frame dropping frame averaging and JPEG

10 Security and Communication Networks

Table2Ex

tractedwatermark(w

ithNCandBE

Rvalues)u

nder

imagep

rocessingattack

Attacks

Foreman

Akiyo

Football

Noattack

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

Gaussianno

ise

NC=090

08BE

R=004

04NC=066

74BE

R=01414

NC=1

BER=0

SaltampPepp

erno

ise

NC=1

BER=0

NC=09238

BER=0303

NC=1

BER=0

Specklen

oise

NC=09479

BER=00202

NC=07870

BER=009

09NC=1

BER=0

Medianfilter

NC=09732

BER=0101

NC=1

BER=0

NC=1

BER=0

Gaussianlowpassfilter

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

Security and Communication Networks 11

Table3Ex

tractedwatermark(w

ithNCandBE

R)un

derframed

ropp

ingattack

Fram

edropp

ingrate

Foreman

Akiyo

Football

50

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

80

NC=08485

BER=006

06NC=07822

BER=0101

NC=07156

BER=01313

12 Security and Communication Networks

Table4Ex

tractedwatermark(w

ithNCandBE

R)un

derframea

veraging

attack

Fram

eaveraging

rate

Foreman

Akiyo

Football

25

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

50

NC=06819

BER=0111

1NC=0800

BER=009

09NC=1

BER=0

Security and Communication Networks 13

Table5Ex

tractedwatermark(w

ithNCandBE

R)un

derframes

wapping

attack

Fram

eSwapping

rate

Foreman

Akiyo

Football

50

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

100

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

14 Security and Communication Networks

Table6Ex

tractedwatermark(w

ithNCandBE

R)un

derv

ideo

compressio

nattack

Foreman

Akiyo

Football

H264compressio

n(512kbps)

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

MPE

G4compressio

n(1500kbp

s)NC=1

BER=0

NC=08708

BER=00505

NC=1

BER=0

MPE

G4compressio

n(100

0kbp

s)NC=1

BER=0

NC=0848

BER=006

06NC=09238

BER=00303

Security and Communication Networks 15

100

90

80

70

60

50

4010 20 30 40 50 60 70 80 90 100

NC

()

JPEG quality factor

10 20 30 40 50 60 70 80 90 100

JPEG quality factor

ForemanAkiyoFootball

ForemanAkiyoFootball

30

25

20

15

10

5

0

BER

()

Figure 7 NC and BER values of the extracted watermarks under the JPEG compression attack

Table 7 Extracted watermark (with NC and BER) under image processing attacks with StirMark benchmark

Attacks NC BERJPEG 10 05677 01515JPEG 30 09479 00202JPEG 60 08485 00606JPEG 90 10000 000003x3 median filter+ JPEG 09732 00101Gaussian filter +JPEG 07681 01010Sharpening +JPEG 05127 02626removed one row one colon +JPEG 05523 02020

compression attacks It is clear that the proposed schemeachieves a better robustness

44 Complexity of the Proposed Algorithm The proposedwatermarking schema has a low complexity because of thefollowing

(i) The algorithm is done in the MR-SVD domain whichhas an overall complexity O(n) where n indicates thesignal length [23]

(ii) To extract fast motion frames we use a simple thresh-old decision method based on motion energy

(iii) The embedding and extraction are not applied to allthe frames of the video but just to fast motion frames

In fact the execution time of the proposed scheme required inwatermark embedding and extraction processes the followingresults watermark embedding 036 s per frame watermarkextraction 019 s per frame The simulation is conducted on

Intel(R) Core i5 180 GHz processor with 4 GB RAMusingMATLAB verR2013b

5 Conclusion

In this paper we proposed a novel blind video watermarkingscheme in fast motion frames using the SVD the MR-SVD the QIM and the Logistic Map encryption In theproposed technique we solved the problem of embedding thewatermark in all frames by choosing only the frames that havebig motion energy to be the host frames suitable to the HSVUsing this approach we have limited the number of frames tobe processed and also assured a higher imperceptibility of thewatermark In addition the combination of the characteris-tics of the SVD MR-SVD QIM and Logistic Map Encryp-tion makes our scheme secure and robust to a variety ofattacks The experimental results confirm that the proposedwatermarking scheme has good imperceptibility with PSNRgreater than 40 dBTheproposed scheme is robust not only toimage processing attacks but also to frame synchronization

16 Security and Communication Networks

100

95

90

85

80

75

NC

()

NC

()

100

95

90

85

80

75

NC

()

10 15 20 25 30 35 40

frame averaging rate

proposed method[20][16][6]

proposed method

proposed method

[20]

[20]

[16][6]

[6]

[18]

10 20 30 40 50 60 70

frame dropping rate

70

65

100

95

90

85

80

75

70

65

60

JPEG Quality factor20 30 40 50 60 70 80 90

Figure 8 Comparison of experimental results between the proposed technique and other schemes under frame averaging attack framedropping attack and JPEG compression attack

and video compression attacks The comparison results withother algorithms related to video watermarking indicate thesuperiority of our scheme

Data Availability

The data used to support the findings of this study are avail-able from the corresponding author upon request

Conflicts of Interest

The authors declare no conflicts of interest

References

[1] J C IngemarM LMiller A B Jeffrey J Fridrich andT KalkerDigital Watermarking and Steganography Morgan KaufmannPublishers Inc 2008

[2] S Katzenbeisser and F Petitcolas ldquoInformation Hiding Tech-niques for Steganography and Digital Watermarkingrdquo ArtechHouse 2000

[3] H Tao L Chongmin J M Zain and A N Abdalla ldquoRobustimagewatermarking theories and techniques A reviewrdquo Journalof Applied Research and Technology vol 12 no 1 pp 122ndash1382014

Security and Communication Networks 17

[4] M Kutter F Jordan and F Bossen ldquoDigital watermarking ofcolor images using amplitude modulationrdquo Journal of ElectronicImaging vol 7 no 2 pp 326ndash332 1998

[5] A Karmakar A Phadikar B S Phadikar and G K Maity ldquoAblind video watermarking scheme resistant to rotation and col-lusion attacksrdquo Journal of King Saud University - Computer andInformation Sciences vol 28 no 2 pp 199ndash210 2016

[6] T R Singh K M Singh and S Roy ldquoVideo watermarkingscheme based on visual cryptography and scene change detec-tionrdquo AEU - International Journal of Electronics and Communi-cations vol 67 no 8 pp 645ndash651 2013

[7] W Kong B Yang D Wu and X Niu ldquoSVD Based Blind VideoWatermarking Algorithmrdquo in Proceedings of the First Interna-tional Conference on Innovative Computing Information andControl - Volume I (ICICICrsquo06) pp 265ndash268 Beijing ChinaSeptember 2006

[8] R Liu and T Tan ldquoAn SVD-based watermarking scheme forprotecting rightful ownershiprdquo IEEE Transactions on Multime-dia vol 4 no 1 pp 121ndash128 2002

[9] O S Faragallah ldquoEfficient video watermarking based on sin-gular value decomposition in the discrete wavelet transformdomainrdquo AEU - International Journal of Electronics and Com-munications vol 67 no 3 pp 189ndash196 2013

[10] K Niu X Yang and L Xiang ldquoHybrid quasi-3D DWTDCTand SVD video watermarkingrdquo in Proceedings of the 2010 IEEEInternational Conference on Software Engineering and ServiceSciences ICSESS 2010 pp 588ndash591 China July 2010

[11] N I Yassin NM Salem andM I El Adawy ldquoQIM blind videowatermarking scheme based on Wavelet transform and prin-cipal component analysisrdquo Alexandria Engineering Journal vol53 no 4 pp 833ndash842 2014

[12] D K Thind and S Jindal ldquoA semi blind DWT-SVD videowatermarkingrdquo inProceedings of the International Conference onInformation and Communication Technologies ICICT 2014 pp1661ndash1667 India December 2014

[13] N Leelavathy E V Prasad and S Kumar ldquoVideo Watermark-ing Techniques A Reviewrdquo International Journal of ComputerApplications vol 104 no 7 pp 24ndash30 2014

[14] F Hartung and M Kutter ldquoMultimedia watermarking techni-quesrdquo Proceedings of the IEEE vol 87 no 7 pp 1079ndash1107 1999

[15] P Rasti S Samiei M Agoyi S Escalera and G AnbarjafarildquoRobust non-blind color video watermarking using QR decom-position and entropy analysisrdquo Journal of Visual Communica-tion and Image Representation vol 38 pp 838ndash847 2016

[16] S M Youssef A A ElFarag and N M Ghatwary ldquoAdaptivevideo watermarking integrating a fuzzy wavelet-based humanvisual system perceptual modelrdquo Multimedia Tools and Appli-cations vol 73 no 3 pp 1545ndash1573 2014

[17] T Tabassum and S M Islam ldquoA digital video watermarkingtechnique based on identical frame extraction in 3-LevelDWTrdquoin Proceedings of the 2012 15th International Conference onComputer and Information Technology (ICCIT) pp 101ndash106Chittagong Bangladesh December 2012

[18] L Agilandeeswari and K Ganesan ldquoA robust color video water-marking scheme based on hybrid embedding techniquesrdquoMul-timedia Tools and Applications vol 75 no 14 pp 8745ndash87802016

[19] X Jiang Q Liu and Q Wu ldquoA new video watermarking algo-rithm based on shot segmentation and block classificationrdquoMultimedia Tools and Applications vol 62 no 3 pp 545ndash5602013

[20] K Chetan and K Raghavendra ldquoDWT Based Blind DigitalVideo Watermarking Scheme for Video Authenticationrdquo Inter-national Journal of Computer Applications vol 4 no 10 pp 19ndash26 2010

[21] Y JiaW Lin and A A Kassim ldquoEstimating just-noticeable dis-tortion for videordquo IEEE Transactions on Circuits and Systems forVideo Technology vol 16 no 7 pp 820ndash829 2006

[22] H C Andrews and C L Patterson ldquoSingular value decomposi-tions and digital image processingrdquo IEEE Transactions on SignalProcessing vol 24 no 1 pp 26ndash53 1976

[23] R Kakarala and P O Ogunbona ldquoSignal analysis using a multi-resolution form of the singular value decompositionrdquo IEEETransactions on Image Processing vol 10 no 5 pp 723ndash7352001

[24] B Chen and G WWornell ldquoQuantization index modulation aclass of provably goodmethods for digital watermarking and in-formation embeddingrdquo IEEE Transactions on Information The-ory vol 47 no 4 pp 1423ndash1443 2001

[25] M Yaghoobi ldquoA New Approach for Image Encryption UsingChaotic Logistic Maprdquo in Proceedings of the 2008 InternationalConference on Advanced Computer Theory and Engineering(ICACTE) pp 585ndash590 Phuket Thailand December 2008

[26] R Dugad K Ratakonda and N Ahuja ldquoRobust video shotchange detectionrdquo in Proceedings of the 1998 IEEE SecondWork-shop on Multimedia Signal Processing pp 376ndash381 RedondoBeach CA USA 1998

[27] A Anees I Hussain A Algarni and M Aslam ldquoA RobustWatermarking Scheme for Online Multimedia Copyright Pro-tection Using New Chaotic Maprdquo Security and CommunicationNetworks vol 2018 2018

[28] R Thanki V Dwivedi and K Borisagar ldquoA hybrid watermark-ing scheme with CS theory for security of multimedia datardquoJournal of King Saud University - Computer and InformationSciences 2017

[29] httpmediaxiphorgvideoderf Accessed on 06 July 2018[30] B G Haskell and A N Netravali Digital Pictures Representa-

tion Compression and Standards Perseus Publishing 1997[31] httpwwwvirtualduborg Accessed on 06 July 2018[32] F A Petitcolas R J Anderson and M G Kuhn ldquoAttacks on

CopyrightMarking Systemsrdquo in Information Hiding vol 1525 ofLecture Notes in Computer Science pp 218ndash238 Springer BerlinHeidelberg Berlin Heidelberg 1998

[33] httpwwwpetitcolasnetwatermarkingstirmark Accessed on19 September 2018

10 Security and Communication Networks

Table2Ex

tractedwatermark(w

ithNCandBE

Rvalues)u

nder

imagep

rocessingattack

Attacks

Foreman

Akiyo

Football

Noattack

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

Gaussianno

ise

NC=090

08BE

R=004

04NC=066

74BE

R=01414

NC=1

BER=0

SaltampPepp

erno

ise

NC=1

BER=0

NC=09238

BER=0303

NC=1

BER=0

Specklen

oise

NC=09479

BER=00202

NC=07870

BER=009

09NC=1

BER=0

Medianfilter

NC=09732

BER=0101

NC=1

BER=0

NC=1

BER=0

Gaussianlowpassfilter

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

Security and Communication Networks 11

Table3Ex

tractedwatermark(w

ithNCandBE

R)un

derframed

ropp

ingattack

Fram

edropp

ingrate

Foreman

Akiyo

Football

50

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

80

NC=08485

BER=006

06NC=07822

BER=0101

NC=07156

BER=01313

12 Security and Communication Networks

Table4Ex

tractedwatermark(w

ithNCandBE

R)un

derframea

veraging

attack

Fram

eaveraging

rate

Foreman

Akiyo

Football

25

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

50

NC=06819

BER=0111

1NC=0800

BER=009

09NC=1

BER=0

Security and Communication Networks 13

Table5Ex

tractedwatermark(w

ithNCandBE

R)un

derframes

wapping

attack

Fram

eSwapping

rate

Foreman

Akiyo

Football

50

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

100

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

14 Security and Communication Networks

Table6Ex

tractedwatermark(w

ithNCandBE

R)un

derv

ideo

compressio

nattack

Foreman

Akiyo

Football

H264compressio

n(512kbps)

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

MPE

G4compressio

n(1500kbp

s)NC=1

BER=0

NC=08708

BER=00505

NC=1

BER=0

MPE

G4compressio

n(100

0kbp

s)NC=1

BER=0

NC=0848

BER=006

06NC=09238

BER=00303

Security and Communication Networks 15

100

90

80

70

60

50

4010 20 30 40 50 60 70 80 90 100

NC

()

JPEG quality factor

10 20 30 40 50 60 70 80 90 100

JPEG quality factor

ForemanAkiyoFootball

ForemanAkiyoFootball

30

25

20

15

10

5

0

BER

()

Figure 7 NC and BER values of the extracted watermarks under the JPEG compression attack

Table 7 Extracted watermark (with NC and BER) under image processing attacks with StirMark benchmark

Attacks NC BERJPEG 10 05677 01515JPEG 30 09479 00202JPEG 60 08485 00606JPEG 90 10000 000003x3 median filter+ JPEG 09732 00101Gaussian filter +JPEG 07681 01010Sharpening +JPEG 05127 02626removed one row one colon +JPEG 05523 02020

compression attacks It is clear that the proposed schemeachieves a better robustness

44 Complexity of the Proposed Algorithm The proposedwatermarking schema has a low complexity because of thefollowing

(i) The algorithm is done in the MR-SVD domain whichhas an overall complexity O(n) where n indicates thesignal length [23]

(ii) To extract fast motion frames we use a simple thresh-old decision method based on motion energy

(iii) The embedding and extraction are not applied to allthe frames of the video but just to fast motion frames

In fact the execution time of the proposed scheme required inwatermark embedding and extraction processes the followingresults watermark embedding 036 s per frame watermarkextraction 019 s per frame The simulation is conducted on

Intel(R) Core i5 180 GHz processor with 4 GB RAMusingMATLAB verR2013b

5 Conclusion

In this paper we proposed a novel blind video watermarkingscheme in fast motion frames using the SVD the MR-SVD the QIM and the Logistic Map encryption In theproposed technique we solved the problem of embedding thewatermark in all frames by choosing only the frames that havebig motion energy to be the host frames suitable to the HSVUsing this approach we have limited the number of frames tobe processed and also assured a higher imperceptibility of thewatermark In addition the combination of the characteris-tics of the SVD MR-SVD QIM and Logistic Map Encryp-tion makes our scheme secure and robust to a variety ofattacks The experimental results confirm that the proposedwatermarking scheme has good imperceptibility with PSNRgreater than 40 dBTheproposed scheme is robust not only toimage processing attacks but also to frame synchronization

16 Security and Communication Networks

100

95

90

85

80

75

NC

()

NC

()

100

95

90

85

80

75

NC

()

10 15 20 25 30 35 40

frame averaging rate

proposed method[20][16][6]

proposed method

proposed method

[20]

[20]

[16][6]

[6]

[18]

10 20 30 40 50 60 70

frame dropping rate

70

65

100

95

90

85

80

75

70

65

60

JPEG Quality factor20 30 40 50 60 70 80 90

Figure 8 Comparison of experimental results between the proposed technique and other schemes under frame averaging attack framedropping attack and JPEG compression attack

and video compression attacks The comparison results withother algorithms related to video watermarking indicate thesuperiority of our scheme

Data Availability

The data used to support the findings of this study are avail-able from the corresponding author upon request

Conflicts of Interest

The authors declare no conflicts of interest

References

[1] J C IngemarM LMiller A B Jeffrey J Fridrich andT KalkerDigital Watermarking and Steganography Morgan KaufmannPublishers Inc 2008

[2] S Katzenbeisser and F Petitcolas ldquoInformation Hiding Tech-niques for Steganography and Digital Watermarkingrdquo ArtechHouse 2000

[3] H Tao L Chongmin J M Zain and A N Abdalla ldquoRobustimagewatermarking theories and techniques A reviewrdquo Journalof Applied Research and Technology vol 12 no 1 pp 122ndash1382014

Security and Communication Networks 17

[4] M Kutter F Jordan and F Bossen ldquoDigital watermarking ofcolor images using amplitude modulationrdquo Journal of ElectronicImaging vol 7 no 2 pp 326ndash332 1998

[5] A Karmakar A Phadikar B S Phadikar and G K Maity ldquoAblind video watermarking scheme resistant to rotation and col-lusion attacksrdquo Journal of King Saud University - Computer andInformation Sciences vol 28 no 2 pp 199ndash210 2016

[6] T R Singh K M Singh and S Roy ldquoVideo watermarkingscheme based on visual cryptography and scene change detec-tionrdquo AEU - International Journal of Electronics and Communi-cations vol 67 no 8 pp 645ndash651 2013

[7] W Kong B Yang D Wu and X Niu ldquoSVD Based Blind VideoWatermarking Algorithmrdquo in Proceedings of the First Interna-tional Conference on Innovative Computing Information andControl - Volume I (ICICICrsquo06) pp 265ndash268 Beijing ChinaSeptember 2006

[8] R Liu and T Tan ldquoAn SVD-based watermarking scheme forprotecting rightful ownershiprdquo IEEE Transactions on Multime-dia vol 4 no 1 pp 121ndash128 2002

[9] O S Faragallah ldquoEfficient video watermarking based on sin-gular value decomposition in the discrete wavelet transformdomainrdquo AEU - International Journal of Electronics and Com-munications vol 67 no 3 pp 189ndash196 2013

[10] K Niu X Yang and L Xiang ldquoHybrid quasi-3D DWTDCTand SVD video watermarkingrdquo in Proceedings of the 2010 IEEEInternational Conference on Software Engineering and ServiceSciences ICSESS 2010 pp 588ndash591 China July 2010

[11] N I Yassin NM Salem andM I El Adawy ldquoQIM blind videowatermarking scheme based on Wavelet transform and prin-cipal component analysisrdquo Alexandria Engineering Journal vol53 no 4 pp 833ndash842 2014

[12] D K Thind and S Jindal ldquoA semi blind DWT-SVD videowatermarkingrdquo inProceedings of the International Conference onInformation and Communication Technologies ICICT 2014 pp1661ndash1667 India December 2014

[13] N Leelavathy E V Prasad and S Kumar ldquoVideo Watermark-ing Techniques A Reviewrdquo International Journal of ComputerApplications vol 104 no 7 pp 24ndash30 2014

[14] F Hartung and M Kutter ldquoMultimedia watermarking techni-quesrdquo Proceedings of the IEEE vol 87 no 7 pp 1079ndash1107 1999

[15] P Rasti S Samiei M Agoyi S Escalera and G AnbarjafarildquoRobust non-blind color video watermarking using QR decom-position and entropy analysisrdquo Journal of Visual Communica-tion and Image Representation vol 38 pp 838ndash847 2016

[16] S M Youssef A A ElFarag and N M Ghatwary ldquoAdaptivevideo watermarking integrating a fuzzy wavelet-based humanvisual system perceptual modelrdquo Multimedia Tools and Appli-cations vol 73 no 3 pp 1545ndash1573 2014

[17] T Tabassum and S M Islam ldquoA digital video watermarkingtechnique based on identical frame extraction in 3-LevelDWTrdquoin Proceedings of the 2012 15th International Conference onComputer and Information Technology (ICCIT) pp 101ndash106Chittagong Bangladesh December 2012

[18] L Agilandeeswari and K Ganesan ldquoA robust color video water-marking scheme based on hybrid embedding techniquesrdquoMul-timedia Tools and Applications vol 75 no 14 pp 8745ndash87802016

[19] X Jiang Q Liu and Q Wu ldquoA new video watermarking algo-rithm based on shot segmentation and block classificationrdquoMultimedia Tools and Applications vol 62 no 3 pp 545ndash5602013

[20] K Chetan and K Raghavendra ldquoDWT Based Blind DigitalVideo Watermarking Scheme for Video Authenticationrdquo Inter-national Journal of Computer Applications vol 4 no 10 pp 19ndash26 2010

[21] Y JiaW Lin and A A Kassim ldquoEstimating just-noticeable dis-tortion for videordquo IEEE Transactions on Circuits and Systems forVideo Technology vol 16 no 7 pp 820ndash829 2006

[22] H C Andrews and C L Patterson ldquoSingular value decomposi-tions and digital image processingrdquo IEEE Transactions on SignalProcessing vol 24 no 1 pp 26ndash53 1976

[23] R Kakarala and P O Ogunbona ldquoSignal analysis using a multi-resolution form of the singular value decompositionrdquo IEEETransactions on Image Processing vol 10 no 5 pp 723ndash7352001

[24] B Chen and G WWornell ldquoQuantization index modulation aclass of provably goodmethods for digital watermarking and in-formation embeddingrdquo IEEE Transactions on Information The-ory vol 47 no 4 pp 1423ndash1443 2001

[25] M Yaghoobi ldquoA New Approach for Image Encryption UsingChaotic Logistic Maprdquo in Proceedings of the 2008 InternationalConference on Advanced Computer Theory and Engineering(ICACTE) pp 585ndash590 Phuket Thailand December 2008

[26] R Dugad K Ratakonda and N Ahuja ldquoRobust video shotchange detectionrdquo in Proceedings of the 1998 IEEE SecondWork-shop on Multimedia Signal Processing pp 376ndash381 RedondoBeach CA USA 1998

[27] A Anees I Hussain A Algarni and M Aslam ldquoA RobustWatermarking Scheme for Online Multimedia Copyright Pro-tection Using New Chaotic Maprdquo Security and CommunicationNetworks vol 2018 2018

[28] R Thanki V Dwivedi and K Borisagar ldquoA hybrid watermark-ing scheme with CS theory for security of multimedia datardquoJournal of King Saud University - Computer and InformationSciences 2017

[29] httpmediaxiphorgvideoderf Accessed on 06 July 2018[30] B G Haskell and A N Netravali Digital Pictures Representa-

tion Compression and Standards Perseus Publishing 1997[31] httpwwwvirtualduborg Accessed on 06 July 2018[32] F A Petitcolas R J Anderson and M G Kuhn ldquoAttacks on

CopyrightMarking Systemsrdquo in Information Hiding vol 1525 ofLecture Notes in Computer Science pp 218ndash238 Springer BerlinHeidelberg Berlin Heidelberg 1998

[33] httpwwwpetitcolasnetwatermarkingstirmark Accessed on19 September 2018

Security and Communication Networks 11

Table3Ex

tractedwatermark(w

ithNCandBE

R)un

derframed

ropp

ingattack

Fram

edropp

ingrate

Foreman

Akiyo

Football

50

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

80

NC=08485

BER=006

06NC=07822

BER=0101

NC=07156

BER=01313

12 Security and Communication Networks

Table4Ex

tractedwatermark(w

ithNCandBE

R)un

derframea

veraging

attack

Fram

eaveraging

rate

Foreman

Akiyo

Football

25

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

50

NC=06819

BER=0111

1NC=0800

BER=009

09NC=1

BER=0

Security and Communication Networks 13

Table5Ex

tractedwatermark(w

ithNCandBE

R)un

derframes

wapping

attack

Fram

eSwapping

rate

Foreman

Akiyo

Football

50

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

100

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

14 Security and Communication Networks

Table6Ex

tractedwatermark(w

ithNCandBE

R)un

derv

ideo

compressio

nattack

Foreman

Akiyo

Football

H264compressio

n(512kbps)

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

MPE

G4compressio

n(1500kbp

s)NC=1

BER=0

NC=08708

BER=00505

NC=1

BER=0

MPE

G4compressio

n(100

0kbp

s)NC=1

BER=0

NC=0848

BER=006

06NC=09238

BER=00303

Security and Communication Networks 15

100

90

80

70

60

50

4010 20 30 40 50 60 70 80 90 100

NC

()

JPEG quality factor

10 20 30 40 50 60 70 80 90 100

JPEG quality factor

ForemanAkiyoFootball

ForemanAkiyoFootball

30

25

20

15

10

5

0

BER

()

Figure 7 NC and BER values of the extracted watermarks under the JPEG compression attack

Table 7 Extracted watermark (with NC and BER) under image processing attacks with StirMark benchmark

Attacks NC BERJPEG 10 05677 01515JPEG 30 09479 00202JPEG 60 08485 00606JPEG 90 10000 000003x3 median filter+ JPEG 09732 00101Gaussian filter +JPEG 07681 01010Sharpening +JPEG 05127 02626removed one row one colon +JPEG 05523 02020

compression attacks It is clear that the proposed schemeachieves a better robustness

44 Complexity of the Proposed Algorithm The proposedwatermarking schema has a low complexity because of thefollowing

(i) The algorithm is done in the MR-SVD domain whichhas an overall complexity O(n) where n indicates thesignal length [23]

(ii) To extract fast motion frames we use a simple thresh-old decision method based on motion energy

(iii) The embedding and extraction are not applied to allthe frames of the video but just to fast motion frames

In fact the execution time of the proposed scheme required inwatermark embedding and extraction processes the followingresults watermark embedding 036 s per frame watermarkextraction 019 s per frame The simulation is conducted on

Intel(R) Core i5 180 GHz processor with 4 GB RAMusingMATLAB verR2013b

5 Conclusion

In this paper we proposed a novel blind video watermarkingscheme in fast motion frames using the SVD the MR-SVD the QIM and the Logistic Map encryption In theproposed technique we solved the problem of embedding thewatermark in all frames by choosing only the frames that havebig motion energy to be the host frames suitable to the HSVUsing this approach we have limited the number of frames tobe processed and also assured a higher imperceptibility of thewatermark In addition the combination of the characteris-tics of the SVD MR-SVD QIM and Logistic Map Encryp-tion makes our scheme secure and robust to a variety ofattacks The experimental results confirm that the proposedwatermarking scheme has good imperceptibility with PSNRgreater than 40 dBTheproposed scheme is robust not only toimage processing attacks but also to frame synchronization

16 Security and Communication Networks

100

95

90

85

80

75

NC

()

NC

()

100

95

90

85

80

75

NC

()

10 15 20 25 30 35 40

frame averaging rate

proposed method[20][16][6]

proposed method

proposed method

[20]

[20]

[16][6]

[6]

[18]

10 20 30 40 50 60 70

frame dropping rate

70

65

100

95

90

85

80

75

70

65

60

JPEG Quality factor20 30 40 50 60 70 80 90

Figure 8 Comparison of experimental results between the proposed technique and other schemes under frame averaging attack framedropping attack and JPEG compression attack

and video compression attacks The comparison results withother algorithms related to video watermarking indicate thesuperiority of our scheme

Data Availability

The data used to support the findings of this study are avail-able from the corresponding author upon request

Conflicts of Interest

The authors declare no conflicts of interest

References

[1] J C IngemarM LMiller A B Jeffrey J Fridrich andT KalkerDigital Watermarking and Steganography Morgan KaufmannPublishers Inc 2008

[2] S Katzenbeisser and F Petitcolas ldquoInformation Hiding Tech-niques for Steganography and Digital Watermarkingrdquo ArtechHouse 2000

[3] H Tao L Chongmin J M Zain and A N Abdalla ldquoRobustimagewatermarking theories and techniques A reviewrdquo Journalof Applied Research and Technology vol 12 no 1 pp 122ndash1382014

Security and Communication Networks 17

[4] M Kutter F Jordan and F Bossen ldquoDigital watermarking ofcolor images using amplitude modulationrdquo Journal of ElectronicImaging vol 7 no 2 pp 326ndash332 1998

[5] A Karmakar A Phadikar B S Phadikar and G K Maity ldquoAblind video watermarking scheme resistant to rotation and col-lusion attacksrdquo Journal of King Saud University - Computer andInformation Sciences vol 28 no 2 pp 199ndash210 2016

[6] T R Singh K M Singh and S Roy ldquoVideo watermarkingscheme based on visual cryptography and scene change detec-tionrdquo AEU - International Journal of Electronics and Communi-cations vol 67 no 8 pp 645ndash651 2013

[7] W Kong B Yang D Wu and X Niu ldquoSVD Based Blind VideoWatermarking Algorithmrdquo in Proceedings of the First Interna-tional Conference on Innovative Computing Information andControl - Volume I (ICICICrsquo06) pp 265ndash268 Beijing ChinaSeptember 2006

[8] R Liu and T Tan ldquoAn SVD-based watermarking scheme forprotecting rightful ownershiprdquo IEEE Transactions on Multime-dia vol 4 no 1 pp 121ndash128 2002

[9] O S Faragallah ldquoEfficient video watermarking based on sin-gular value decomposition in the discrete wavelet transformdomainrdquo AEU - International Journal of Electronics and Com-munications vol 67 no 3 pp 189ndash196 2013

[10] K Niu X Yang and L Xiang ldquoHybrid quasi-3D DWTDCTand SVD video watermarkingrdquo in Proceedings of the 2010 IEEEInternational Conference on Software Engineering and ServiceSciences ICSESS 2010 pp 588ndash591 China July 2010

[11] N I Yassin NM Salem andM I El Adawy ldquoQIM blind videowatermarking scheme based on Wavelet transform and prin-cipal component analysisrdquo Alexandria Engineering Journal vol53 no 4 pp 833ndash842 2014

[12] D K Thind and S Jindal ldquoA semi blind DWT-SVD videowatermarkingrdquo inProceedings of the International Conference onInformation and Communication Technologies ICICT 2014 pp1661ndash1667 India December 2014

[13] N Leelavathy E V Prasad and S Kumar ldquoVideo Watermark-ing Techniques A Reviewrdquo International Journal of ComputerApplications vol 104 no 7 pp 24ndash30 2014

[14] F Hartung and M Kutter ldquoMultimedia watermarking techni-quesrdquo Proceedings of the IEEE vol 87 no 7 pp 1079ndash1107 1999

[15] P Rasti S Samiei M Agoyi S Escalera and G AnbarjafarildquoRobust non-blind color video watermarking using QR decom-position and entropy analysisrdquo Journal of Visual Communica-tion and Image Representation vol 38 pp 838ndash847 2016

[16] S M Youssef A A ElFarag and N M Ghatwary ldquoAdaptivevideo watermarking integrating a fuzzy wavelet-based humanvisual system perceptual modelrdquo Multimedia Tools and Appli-cations vol 73 no 3 pp 1545ndash1573 2014

[17] T Tabassum and S M Islam ldquoA digital video watermarkingtechnique based on identical frame extraction in 3-LevelDWTrdquoin Proceedings of the 2012 15th International Conference onComputer and Information Technology (ICCIT) pp 101ndash106Chittagong Bangladesh December 2012

[18] L Agilandeeswari and K Ganesan ldquoA robust color video water-marking scheme based on hybrid embedding techniquesrdquoMul-timedia Tools and Applications vol 75 no 14 pp 8745ndash87802016

[19] X Jiang Q Liu and Q Wu ldquoA new video watermarking algo-rithm based on shot segmentation and block classificationrdquoMultimedia Tools and Applications vol 62 no 3 pp 545ndash5602013

[20] K Chetan and K Raghavendra ldquoDWT Based Blind DigitalVideo Watermarking Scheme for Video Authenticationrdquo Inter-national Journal of Computer Applications vol 4 no 10 pp 19ndash26 2010

[21] Y JiaW Lin and A A Kassim ldquoEstimating just-noticeable dis-tortion for videordquo IEEE Transactions on Circuits and Systems forVideo Technology vol 16 no 7 pp 820ndash829 2006

[22] H C Andrews and C L Patterson ldquoSingular value decomposi-tions and digital image processingrdquo IEEE Transactions on SignalProcessing vol 24 no 1 pp 26ndash53 1976

[23] R Kakarala and P O Ogunbona ldquoSignal analysis using a multi-resolution form of the singular value decompositionrdquo IEEETransactions on Image Processing vol 10 no 5 pp 723ndash7352001

[24] B Chen and G WWornell ldquoQuantization index modulation aclass of provably goodmethods for digital watermarking and in-formation embeddingrdquo IEEE Transactions on Information The-ory vol 47 no 4 pp 1423ndash1443 2001

[25] M Yaghoobi ldquoA New Approach for Image Encryption UsingChaotic Logistic Maprdquo in Proceedings of the 2008 InternationalConference on Advanced Computer Theory and Engineering(ICACTE) pp 585ndash590 Phuket Thailand December 2008

[26] R Dugad K Ratakonda and N Ahuja ldquoRobust video shotchange detectionrdquo in Proceedings of the 1998 IEEE SecondWork-shop on Multimedia Signal Processing pp 376ndash381 RedondoBeach CA USA 1998

[27] A Anees I Hussain A Algarni and M Aslam ldquoA RobustWatermarking Scheme for Online Multimedia Copyright Pro-tection Using New Chaotic Maprdquo Security and CommunicationNetworks vol 2018 2018

[28] R Thanki V Dwivedi and K Borisagar ldquoA hybrid watermark-ing scheme with CS theory for security of multimedia datardquoJournal of King Saud University - Computer and InformationSciences 2017

[29] httpmediaxiphorgvideoderf Accessed on 06 July 2018[30] B G Haskell and A N Netravali Digital Pictures Representa-

tion Compression and Standards Perseus Publishing 1997[31] httpwwwvirtualduborg Accessed on 06 July 2018[32] F A Petitcolas R J Anderson and M G Kuhn ldquoAttacks on

CopyrightMarking Systemsrdquo in Information Hiding vol 1525 ofLecture Notes in Computer Science pp 218ndash238 Springer BerlinHeidelberg Berlin Heidelberg 1998

[33] httpwwwpetitcolasnetwatermarkingstirmark Accessed on19 September 2018

12 Security and Communication Networks

Table4Ex

tractedwatermark(w

ithNCandBE

R)un

derframea

veraging

attack

Fram

eaveraging

rate

Foreman

Akiyo

Football

25

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

50

NC=06819

BER=0111

1NC=0800

BER=009

09NC=1

BER=0

Security and Communication Networks 13

Table5Ex

tractedwatermark(w

ithNCandBE

R)un

derframes

wapping

attack

Fram

eSwapping

rate

Foreman

Akiyo

Football

50

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

100

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

14 Security and Communication Networks

Table6Ex

tractedwatermark(w

ithNCandBE

R)un

derv

ideo

compressio

nattack

Foreman

Akiyo

Football

H264compressio

n(512kbps)

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

MPE

G4compressio

n(1500kbp

s)NC=1

BER=0

NC=08708

BER=00505

NC=1

BER=0

MPE

G4compressio

n(100

0kbp

s)NC=1

BER=0

NC=0848

BER=006

06NC=09238

BER=00303

Security and Communication Networks 15

100

90

80

70

60

50

4010 20 30 40 50 60 70 80 90 100

NC

()

JPEG quality factor

10 20 30 40 50 60 70 80 90 100

JPEG quality factor

ForemanAkiyoFootball

ForemanAkiyoFootball

30

25

20

15

10

5

0

BER

()

Figure 7 NC and BER values of the extracted watermarks under the JPEG compression attack

Table 7 Extracted watermark (with NC and BER) under image processing attacks with StirMark benchmark

Attacks NC BERJPEG 10 05677 01515JPEG 30 09479 00202JPEG 60 08485 00606JPEG 90 10000 000003x3 median filter+ JPEG 09732 00101Gaussian filter +JPEG 07681 01010Sharpening +JPEG 05127 02626removed one row one colon +JPEG 05523 02020

compression attacks It is clear that the proposed schemeachieves a better robustness

44 Complexity of the Proposed Algorithm The proposedwatermarking schema has a low complexity because of thefollowing

(i) The algorithm is done in the MR-SVD domain whichhas an overall complexity O(n) where n indicates thesignal length [23]

(ii) To extract fast motion frames we use a simple thresh-old decision method based on motion energy

(iii) The embedding and extraction are not applied to allthe frames of the video but just to fast motion frames

In fact the execution time of the proposed scheme required inwatermark embedding and extraction processes the followingresults watermark embedding 036 s per frame watermarkextraction 019 s per frame The simulation is conducted on

Intel(R) Core i5 180 GHz processor with 4 GB RAMusingMATLAB verR2013b

5 Conclusion

In this paper we proposed a novel blind video watermarkingscheme in fast motion frames using the SVD the MR-SVD the QIM and the Logistic Map encryption In theproposed technique we solved the problem of embedding thewatermark in all frames by choosing only the frames that havebig motion energy to be the host frames suitable to the HSVUsing this approach we have limited the number of frames tobe processed and also assured a higher imperceptibility of thewatermark In addition the combination of the characteris-tics of the SVD MR-SVD QIM and Logistic Map Encryp-tion makes our scheme secure and robust to a variety ofattacks The experimental results confirm that the proposedwatermarking scheme has good imperceptibility with PSNRgreater than 40 dBTheproposed scheme is robust not only toimage processing attacks but also to frame synchronization

16 Security and Communication Networks

100

95

90

85

80

75

NC

()

NC

()

100

95

90

85

80

75

NC

()

10 15 20 25 30 35 40

frame averaging rate

proposed method[20][16][6]

proposed method

proposed method

[20]

[20]

[16][6]

[6]

[18]

10 20 30 40 50 60 70

frame dropping rate

70

65

100

95

90

85

80

75

70

65

60

JPEG Quality factor20 30 40 50 60 70 80 90

Figure 8 Comparison of experimental results between the proposed technique and other schemes under frame averaging attack framedropping attack and JPEG compression attack

and video compression attacks The comparison results withother algorithms related to video watermarking indicate thesuperiority of our scheme

Data Availability

The data used to support the findings of this study are avail-able from the corresponding author upon request

Conflicts of Interest

The authors declare no conflicts of interest

References

[1] J C IngemarM LMiller A B Jeffrey J Fridrich andT KalkerDigital Watermarking and Steganography Morgan KaufmannPublishers Inc 2008

[2] S Katzenbeisser and F Petitcolas ldquoInformation Hiding Tech-niques for Steganography and Digital Watermarkingrdquo ArtechHouse 2000

[3] H Tao L Chongmin J M Zain and A N Abdalla ldquoRobustimagewatermarking theories and techniques A reviewrdquo Journalof Applied Research and Technology vol 12 no 1 pp 122ndash1382014

Security and Communication Networks 17

[4] M Kutter F Jordan and F Bossen ldquoDigital watermarking ofcolor images using amplitude modulationrdquo Journal of ElectronicImaging vol 7 no 2 pp 326ndash332 1998

[5] A Karmakar A Phadikar B S Phadikar and G K Maity ldquoAblind video watermarking scheme resistant to rotation and col-lusion attacksrdquo Journal of King Saud University - Computer andInformation Sciences vol 28 no 2 pp 199ndash210 2016

[6] T R Singh K M Singh and S Roy ldquoVideo watermarkingscheme based on visual cryptography and scene change detec-tionrdquo AEU - International Journal of Electronics and Communi-cations vol 67 no 8 pp 645ndash651 2013

[7] W Kong B Yang D Wu and X Niu ldquoSVD Based Blind VideoWatermarking Algorithmrdquo in Proceedings of the First Interna-tional Conference on Innovative Computing Information andControl - Volume I (ICICICrsquo06) pp 265ndash268 Beijing ChinaSeptember 2006

[8] R Liu and T Tan ldquoAn SVD-based watermarking scheme forprotecting rightful ownershiprdquo IEEE Transactions on Multime-dia vol 4 no 1 pp 121ndash128 2002

[9] O S Faragallah ldquoEfficient video watermarking based on sin-gular value decomposition in the discrete wavelet transformdomainrdquo AEU - International Journal of Electronics and Com-munications vol 67 no 3 pp 189ndash196 2013

[10] K Niu X Yang and L Xiang ldquoHybrid quasi-3D DWTDCTand SVD video watermarkingrdquo in Proceedings of the 2010 IEEEInternational Conference on Software Engineering and ServiceSciences ICSESS 2010 pp 588ndash591 China July 2010

[11] N I Yassin NM Salem andM I El Adawy ldquoQIM blind videowatermarking scheme based on Wavelet transform and prin-cipal component analysisrdquo Alexandria Engineering Journal vol53 no 4 pp 833ndash842 2014

[12] D K Thind and S Jindal ldquoA semi blind DWT-SVD videowatermarkingrdquo inProceedings of the International Conference onInformation and Communication Technologies ICICT 2014 pp1661ndash1667 India December 2014

[13] N Leelavathy E V Prasad and S Kumar ldquoVideo Watermark-ing Techniques A Reviewrdquo International Journal of ComputerApplications vol 104 no 7 pp 24ndash30 2014

[14] F Hartung and M Kutter ldquoMultimedia watermarking techni-quesrdquo Proceedings of the IEEE vol 87 no 7 pp 1079ndash1107 1999

[15] P Rasti S Samiei M Agoyi S Escalera and G AnbarjafarildquoRobust non-blind color video watermarking using QR decom-position and entropy analysisrdquo Journal of Visual Communica-tion and Image Representation vol 38 pp 838ndash847 2016

[16] S M Youssef A A ElFarag and N M Ghatwary ldquoAdaptivevideo watermarking integrating a fuzzy wavelet-based humanvisual system perceptual modelrdquo Multimedia Tools and Appli-cations vol 73 no 3 pp 1545ndash1573 2014

[17] T Tabassum and S M Islam ldquoA digital video watermarkingtechnique based on identical frame extraction in 3-LevelDWTrdquoin Proceedings of the 2012 15th International Conference onComputer and Information Technology (ICCIT) pp 101ndash106Chittagong Bangladesh December 2012

[18] L Agilandeeswari and K Ganesan ldquoA robust color video water-marking scheme based on hybrid embedding techniquesrdquoMul-timedia Tools and Applications vol 75 no 14 pp 8745ndash87802016

[19] X Jiang Q Liu and Q Wu ldquoA new video watermarking algo-rithm based on shot segmentation and block classificationrdquoMultimedia Tools and Applications vol 62 no 3 pp 545ndash5602013

[20] K Chetan and K Raghavendra ldquoDWT Based Blind DigitalVideo Watermarking Scheme for Video Authenticationrdquo Inter-national Journal of Computer Applications vol 4 no 10 pp 19ndash26 2010

[21] Y JiaW Lin and A A Kassim ldquoEstimating just-noticeable dis-tortion for videordquo IEEE Transactions on Circuits and Systems forVideo Technology vol 16 no 7 pp 820ndash829 2006

[22] H C Andrews and C L Patterson ldquoSingular value decomposi-tions and digital image processingrdquo IEEE Transactions on SignalProcessing vol 24 no 1 pp 26ndash53 1976

[23] R Kakarala and P O Ogunbona ldquoSignal analysis using a multi-resolution form of the singular value decompositionrdquo IEEETransactions on Image Processing vol 10 no 5 pp 723ndash7352001

[24] B Chen and G WWornell ldquoQuantization index modulation aclass of provably goodmethods for digital watermarking and in-formation embeddingrdquo IEEE Transactions on Information The-ory vol 47 no 4 pp 1423ndash1443 2001

[25] M Yaghoobi ldquoA New Approach for Image Encryption UsingChaotic Logistic Maprdquo in Proceedings of the 2008 InternationalConference on Advanced Computer Theory and Engineering(ICACTE) pp 585ndash590 Phuket Thailand December 2008

[26] R Dugad K Ratakonda and N Ahuja ldquoRobust video shotchange detectionrdquo in Proceedings of the 1998 IEEE SecondWork-shop on Multimedia Signal Processing pp 376ndash381 RedondoBeach CA USA 1998

[27] A Anees I Hussain A Algarni and M Aslam ldquoA RobustWatermarking Scheme for Online Multimedia Copyright Pro-tection Using New Chaotic Maprdquo Security and CommunicationNetworks vol 2018 2018

[28] R Thanki V Dwivedi and K Borisagar ldquoA hybrid watermark-ing scheme with CS theory for security of multimedia datardquoJournal of King Saud University - Computer and InformationSciences 2017

[29] httpmediaxiphorgvideoderf Accessed on 06 July 2018[30] B G Haskell and A N Netravali Digital Pictures Representa-

tion Compression and Standards Perseus Publishing 1997[31] httpwwwvirtualduborg Accessed on 06 July 2018[32] F A Petitcolas R J Anderson and M G Kuhn ldquoAttacks on

CopyrightMarking Systemsrdquo in Information Hiding vol 1525 ofLecture Notes in Computer Science pp 218ndash238 Springer BerlinHeidelberg Berlin Heidelberg 1998

[33] httpwwwpetitcolasnetwatermarkingstirmark Accessed on19 September 2018

Security and Communication Networks 13

Table5Ex

tractedwatermark(w

ithNCandBE

R)un

derframes

wapping

attack

Fram

eSwapping

rate

Foreman

Akiyo

Football

50

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

100

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

14 Security and Communication Networks

Table6Ex

tractedwatermark(w

ithNCandBE

R)un

derv

ideo

compressio

nattack

Foreman

Akiyo

Football

H264compressio

n(512kbps)

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

MPE

G4compressio

n(1500kbp

s)NC=1

BER=0

NC=08708

BER=00505

NC=1

BER=0

MPE

G4compressio

n(100

0kbp

s)NC=1

BER=0

NC=0848

BER=006

06NC=09238

BER=00303

Security and Communication Networks 15

100

90

80

70

60

50

4010 20 30 40 50 60 70 80 90 100

NC

()

JPEG quality factor

10 20 30 40 50 60 70 80 90 100

JPEG quality factor

ForemanAkiyoFootball

ForemanAkiyoFootball

30

25

20

15

10

5

0

BER

()

Figure 7 NC and BER values of the extracted watermarks under the JPEG compression attack

Table 7 Extracted watermark (with NC and BER) under image processing attacks with StirMark benchmark

Attacks NC BERJPEG 10 05677 01515JPEG 30 09479 00202JPEG 60 08485 00606JPEG 90 10000 000003x3 median filter+ JPEG 09732 00101Gaussian filter +JPEG 07681 01010Sharpening +JPEG 05127 02626removed one row one colon +JPEG 05523 02020

compression attacks It is clear that the proposed schemeachieves a better robustness

44 Complexity of the Proposed Algorithm The proposedwatermarking schema has a low complexity because of thefollowing

(i) The algorithm is done in the MR-SVD domain whichhas an overall complexity O(n) where n indicates thesignal length [23]

(ii) To extract fast motion frames we use a simple thresh-old decision method based on motion energy

(iii) The embedding and extraction are not applied to allthe frames of the video but just to fast motion frames

In fact the execution time of the proposed scheme required inwatermark embedding and extraction processes the followingresults watermark embedding 036 s per frame watermarkextraction 019 s per frame The simulation is conducted on

Intel(R) Core i5 180 GHz processor with 4 GB RAMusingMATLAB verR2013b

5 Conclusion

In this paper we proposed a novel blind video watermarkingscheme in fast motion frames using the SVD the MR-SVD the QIM and the Logistic Map encryption In theproposed technique we solved the problem of embedding thewatermark in all frames by choosing only the frames that havebig motion energy to be the host frames suitable to the HSVUsing this approach we have limited the number of frames tobe processed and also assured a higher imperceptibility of thewatermark In addition the combination of the characteris-tics of the SVD MR-SVD QIM and Logistic Map Encryp-tion makes our scheme secure and robust to a variety ofattacks The experimental results confirm that the proposedwatermarking scheme has good imperceptibility with PSNRgreater than 40 dBTheproposed scheme is robust not only toimage processing attacks but also to frame synchronization

16 Security and Communication Networks

100

95

90

85

80

75

NC

()

NC

()

100

95

90

85

80

75

NC

()

10 15 20 25 30 35 40

frame averaging rate

proposed method[20][16][6]

proposed method

proposed method

[20]

[20]

[16][6]

[6]

[18]

10 20 30 40 50 60 70

frame dropping rate

70

65

100

95

90

85

80

75

70

65

60

JPEG Quality factor20 30 40 50 60 70 80 90

Figure 8 Comparison of experimental results between the proposed technique and other schemes under frame averaging attack framedropping attack and JPEG compression attack

and video compression attacks The comparison results withother algorithms related to video watermarking indicate thesuperiority of our scheme

Data Availability

The data used to support the findings of this study are avail-able from the corresponding author upon request

Conflicts of Interest

The authors declare no conflicts of interest

References

[1] J C IngemarM LMiller A B Jeffrey J Fridrich andT KalkerDigital Watermarking and Steganography Morgan KaufmannPublishers Inc 2008

[2] S Katzenbeisser and F Petitcolas ldquoInformation Hiding Tech-niques for Steganography and Digital Watermarkingrdquo ArtechHouse 2000

[3] H Tao L Chongmin J M Zain and A N Abdalla ldquoRobustimagewatermarking theories and techniques A reviewrdquo Journalof Applied Research and Technology vol 12 no 1 pp 122ndash1382014

Security and Communication Networks 17

[4] M Kutter F Jordan and F Bossen ldquoDigital watermarking ofcolor images using amplitude modulationrdquo Journal of ElectronicImaging vol 7 no 2 pp 326ndash332 1998

[5] A Karmakar A Phadikar B S Phadikar and G K Maity ldquoAblind video watermarking scheme resistant to rotation and col-lusion attacksrdquo Journal of King Saud University - Computer andInformation Sciences vol 28 no 2 pp 199ndash210 2016

[6] T R Singh K M Singh and S Roy ldquoVideo watermarkingscheme based on visual cryptography and scene change detec-tionrdquo AEU - International Journal of Electronics and Communi-cations vol 67 no 8 pp 645ndash651 2013

[7] W Kong B Yang D Wu and X Niu ldquoSVD Based Blind VideoWatermarking Algorithmrdquo in Proceedings of the First Interna-tional Conference on Innovative Computing Information andControl - Volume I (ICICICrsquo06) pp 265ndash268 Beijing ChinaSeptember 2006

[8] R Liu and T Tan ldquoAn SVD-based watermarking scheme forprotecting rightful ownershiprdquo IEEE Transactions on Multime-dia vol 4 no 1 pp 121ndash128 2002

[9] O S Faragallah ldquoEfficient video watermarking based on sin-gular value decomposition in the discrete wavelet transformdomainrdquo AEU - International Journal of Electronics and Com-munications vol 67 no 3 pp 189ndash196 2013

[10] K Niu X Yang and L Xiang ldquoHybrid quasi-3D DWTDCTand SVD video watermarkingrdquo in Proceedings of the 2010 IEEEInternational Conference on Software Engineering and ServiceSciences ICSESS 2010 pp 588ndash591 China July 2010

[11] N I Yassin NM Salem andM I El Adawy ldquoQIM blind videowatermarking scheme based on Wavelet transform and prin-cipal component analysisrdquo Alexandria Engineering Journal vol53 no 4 pp 833ndash842 2014

[12] D K Thind and S Jindal ldquoA semi blind DWT-SVD videowatermarkingrdquo inProceedings of the International Conference onInformation and Communication Technologies ICICT 2014 pp1661ndash1667 India December 2014

[13] N Leelavathy E V Prasad and S Kumar ldquoVideo Watermark-ing Techniques A Reviewrdquo International Journal of ComputerApplications vol 104 no 7 pp 24ndash30 2014

[14] F Hartung and M Kutter ldquoMultimedia watermarking techni-quesrdquo Proceedings of the IEEE vol 87 no 7 pp 1079ndash1107 1999

[15] P Rasti S Samiei M Agoyi S Escalera and G AnbarjafarildquoRobust non-blind color video watermarking using QR decom-position and entropy analysisrdquo Journal of Visual Communica-tion and Image Representation vol 38 pp 838ndash847 2016

[16] S M Youssef A A ElFarag and N M Ghatwary ldquoAdaptivevideo watermarking integrating a fuzzy wavelet-based humanvisual system perceptual modelrdquo Multimedia Tools and Appli-cations vol 73 no 3 pp 1545ndash1573 2014

[17] T Tabassum and S M Islam ldquoA digital video watermarkingtechnique based on identical frame extraction in 3-LevelDWTrdquoin Proceedings of the 2012 15th International Conference onComputer and Information Technology (ICCIT) pp 101ndash106Chittagong Bangladesh December 2012

[18] L Agilandeeswari and K Ganesan ldquoA robust color video water-marking scheme based on hybrid embedding techniquesrdquoMul-timedia Tools and Applications vol 75 no 14 pp 8745ndash87802016

[19] X Jiang Q Liu and Q Wu ldquoA new video watermarking algo-rithm based on shot segmentation and block classificationrdquoMultimedia Tools and Applications vol 62 no 3 pp 545ndash5602013

[20] K Chetan and K Raghavendra ldquoDWT Based Blind DigitalVideo Watermarking Scheme for Video Authenticationrdquo Inter-national Journal of Computer Applications vol 4 no 10 pp 19ndash26 2010

[21] Y JiaW Lin and A A Kassim ldquoEstimating just-noticeable dis-tortion for videordquo IEEE Transactions on Circuits and Systems forVideo Technology vol 16 no 7 pp 820ndash829 2006

[22] H C Andrews and C L Patterson ldquoSingular value decomposi-tions and digital image processingrdquo IEEE Transactions on SignalProcessing vol 24 no 1 pp 26ndash53 1976

[23] R Kakarala and P O Ogunbona ldquoSignal analysis using a multi-resolution form of the singular value decompositionrdquo IEEETransactions on Image Processing vol 10 no 5 pp 723ndash7352001

[24] B Chen and G WWornell ldquoQuantization index modulation aclass of provably goodmethods for digital watermarking and in-formation embeddingrdquo IEEE Transactions on Information The-ory vol 47 no 4 pp 1423ndash1443 2001

[25] M Yaghoobi ldquoA New Approach for Image Encryption UsingChaotic Logistic Maprdquo in Proceedings of the 2008 InternationalConference on Advanced Computer Theory and Engineering(ICACTE) pp 585ndash590 Phuket Thailand December 2008

[26] R Dugad K Ratakonda and N Ahuja ldquoRobust video shotchange detectionrdquo in Proceedings of the 1998 IEEE SecondWork-shop on Multimedia Signal Processing pp 376ndash381 RedondoBeach CA USA 1998

[27] A Anees I Hussain A Algarni and M Aslam ldquoA RobustWatermarking Scheme for Online Multimedia Copyright Pro-tection Using New Chaotic Maprdquo Security and CommunicationNetworks vol 2018 2018

[28] R Thanki V Dwivedi and K Borisagar ldquoA hybrid watermark-ing scheme with CS theory for security of multimedia datardquoJournal of King Saud University - Computer and InformationSciences 2017

[29] httpmediaxiphorgvideoderf Accessed on 06 July 2018[30] B G Haskell and A N Netravali Digital Pictures Representa-

tion Compression and Standards Perseus Publishing 1997[31] httpwwwvirtualduborg Accessed on 06 July 2018[32] F A Petitcolas R J Anderson and M G Kuhn ldquoAttacks on

CopyrightMarking Systemsrdquo in Information Hiding vol 1525 ofLecture Notes in Computer Science pp 218ndash238 Springer BerlinHeidelberg Berlin Heidelberg 1998

[33] httpwwwpetitcolasnetwatermarkingstirmark Accessed on19 September 2018

14 Security and Communication Networks

Table6Ex

tractedwatermark(w

ithNCandBE

R)un

derv

ideo

compressio

nattack

Foreman

Akiyo

Football

H264compressio

n(512kbps)

NC=1

BER=0

NC=1

BER=0

NC=1

BER=0

MPE

G4compressio

n(1500kbp

s)NC=1

BER=0

NC=08708

BER=00505

NC=1

BER=0

MPE

G4compressio

n(100

0kbp

s)NC=1

BER=0

NC=0848

BER=006

06NC=09238

BER=00303

Security and Communication Networks 15

100

90

80

70

60

50

4010 20 30 40 50 60 70 80 90 100

NC

()

JPEG quality factor

10 20 30 40 50 60 70 80 90 100

JPEG quality factor

ForemanAkiyoFootball

ForemanAkiyoFootball

30

25

20

15

10

5

0

BER

()

Figure 7 NC and BER values of the extracted watermarks under the JPEG compression attack

Table 7 Extracted watermark (with NC and BER) under image processing attacks with StirMark benchmark

Attacks NC BERJPEG 10 05677 01515JPEG 30 09479 00202JPEG 60 08485 00606JPEG 90 10000 000003x3 median filter+ JPEG 09732 00101Gaussian filter +JPEG 07681 01010Sharpening +JPEG 05127 02626removed one row one colon +JPEG 05523 02020

compression attacks It is clear that the proposed schemeachieves a better robustness

44 Complexity of the Proposed Algorithm The proposedwatermarking schema has a low complexity because of thefollowing

(i) The algorithm is done in the MR-SVD domain whichhas an overall complexity O(n) where n indicates thesignal length [23]

(ii) To extract fast motion frames we use a simple thresh-old decision method based on motion energy

(iii) The embedding and extraction are not applied to allthe frames of the video but just to fast motion frames

In fact the execution time of the proposed scheme required inwatermark embedding and extraction processes the followingresults watermark embedding 036 s per frame watermarkextraction 019 s per frame The simulation is conducted on

Intel(R) Core i5 180 GHz processor with 4 GB RAMusingMATLAB verR2013b

5 Conclusion

In this paper we proposed a novel blind video watermarkingscheme in fast motion frames using the SVD the MR-SVD the QIM and the Logistic Map encryption In theproposed technique we solved the problem of embedding thewatermark in all frames by choosing only the frames that havebig motion energy to be the host frames suitable to the HSVUsing this approach we have limited the number of frames tobe processed and also assured a higher imperceptibility of thewatermark In addition the combination of the characteris-tics of the SVD MR-SVD QIM and Logistic Map Encryp-tion makes our scheme secure and robust to a variety ofattacks The experimental results confirm that the proposedwatermarking scheme has good imperceptibility with PSNRgreater than 40 dBTheproposed scheme is robust not only toimage processing attacks but also to frame synchronization

16 Security and Communication Networks

100

95

90

85

80

75

NC

()

NC

()

100

95

90

85

80

75

NC

()

10 15 20 25 30 35 40

frame averaging rate

proposed method[20][16][6]

proposed method

proposed method

[20]

[20]

[16][6]

[6]

[18]

10 20 30 40 50 60 70

frame dropping rate

70

65

100

95

90

85

80

75

70

65

60

JPEG Quality factor20 30 40 50 60 70 80 90

Figure 8 Comparison of experimental results between the proposed technique and other schemes under frame averaging attack framedropping attack and JPEG compression attack

and video compression attacks The comparison results withother algorithms related to video watermarking indicate thesuperiority of our scheme

Data Availability

The data used to support the findings of this study are avail-able from the corresponding author upon request

Conflicts of Interest

The authors declare no conflicts of interest

References

[1] J C IngemarM LMiller A B Jeffrey J Fridrich andT KalkerDigital Watermarking and Steganography Morgan KaufmannPublishers Inc 2008

[2] S Katzenbeisser and F Petitcolas ldquoInformation Hiding Tech-niques for Steganography and Digital Watermarkingrdquo ArtechHouse 2000

[3] H Tao L Chongmin J M Zain and A N Abdalla ldquoRobustimagewatermarking theories and techniques A reviewrdquo Journalof Applied Research and Technology vol 12 no 1 pp 122ndash1382014

Security and Communication Networks 17

[4] M Kutter F Jordan and F Bossen ldquoDigital watermarking ofcolor images using amplitude modulationrdquo Journal of ElectronicImaging vol 7 no 2 pp 326ndash332 1998

[5] A Karmakar A Phadikar B S Phadikar and G K Maity ldquoAblind video watermarking scheme resistant to rotation and col-lusion attacksrdquo Journal of King Saud University - Computer andInformation Sciences vol 28 no 2 pp 199ndash210 2016

[6] T R Singh K M Singh and S Roy ldquoVideo watermarkingscheme based on visual cryptography and scene change detec-tionrdquo AEU - International Journal of Electronics and Communi-cations vol 67 no 8 pp 645ndash651 2013

[7] W Kong B Yang D Wu and X Niu ldquoSVD Based Blind VideoWatermarking Algorithmrdquo in Proceedings of the First Interna-tional Conference on Innovative Computing Information andControl - Volume I (ICICICrsquo06) pp 265ndash268 Beijing ChinaSeptember 2006

[8] R Liu and T Tan ldquoAn SVD-based watermarking scheme forprotecting rightful ownershiprdquo IEEE Transactions on Multime-dia vol 4 no 1 pp 121ndash128 2002

[9] O S Faragallah ldquoEfficient video watermarking based on sin-gular value decomposition in the discrete wavelet transformdomainrdquo AEU - International Journal of Electronics and Com-munications vol 67 no 3 pp 189ndash196 2013

[10] K Niu X Yang and L Xiang ldquoHybrid quasi-3D DWTDCTand SVD video watermarkingrdquo in Proceedings of the 2010 IEEEInternational Conference on Software Engineering and ServiceSciences ICSESS 2010 pp 588ndash591 China July 2010

[11] N I Yassin NM Salem andM I El Adawy ldquoQIM blind videowatermarking scheme based on Wavelet transform and prin-cipal component analysisrdquo Alexandria Engineering Journal vol53 no 4 pp 833ndash842 2014

[12] D K Thind and S Jindal ldquoA semi blind DWT-SVD videowatermarkingrdquo inProceedings of the International Conference onInformation and Communication Technologies ICICT 2014 pp1661ndash1667 India December 2014

[13] N Leelavathy E V Prasad and S Kumar ldquoVideo Watermark-ing Techniques A Reviewrdquo International Journal of ComputerApplications vol 104 no 7 pp 24ndash30 2014

[14] F Hartung and M Kutter ldquoMultimedia watermarking techni-quesrdquo Proceedings of the IEEE vol 87 no 7 pp 1079ndash1107 1999

[15] P Rasti S Samiei M Agoyi S Escalera and G AnbarjafarildquoRobust non-blind color video watermarking using QR decom-position and entropy analysisrdquo Journal of Visual Communica-tion and Image Representation vol 38 pp 838ndash847 2016

[16] S M Youssef A A ElFarag and N M Ghatwary ldquoAdaptivevideo watermarking integrating a fuzzy wavelet-based humanvisual system perceptual modelrdquo Multimedia Tools and Appli-cations vol 73 no 3 pp 1545ndash1573 2014

[17] T Tabassum and S M Islam ldquoA digital video watermarkingtechnique based on identical frame extraction in 3-LevelDWTrdquoin Proceedings of the 2012 15th International Conference onComputer and Information Technology (ICCIT) pp 101ndash106Chittagong Bangladesh December 2012

[18] L Agilandeeswari and K Ganesan ldquoA robust color video water-marking scheme based on hybrid embedding techniquesrdquoMul-timedia Tools and Applications vol 75 no 14 pp 8745ndash87802016

[19] X Jiang Q Liu and Q Wu ldquoA new video watermarking algo-rithm based on shot segmentation and block classificationrdquoMultimedia Tools and Applications vol 62 no 3 pp 545ndash5602013

[20] K Chetan and K Raghavendra ldquoDWT Based Blind DigitalVideo Watermarking Scheme for Video Authenticationrdquo Inter-national Journal of Computer Applications vol 4 no 10 pp 19ndash26 2010

[21] Y JiaW Lin and A A Kassim ldquoEstimating just-noticeable dis-tortion for videordquo IEEE Transactions on Circuits and Systems forVideo Technology vol 16 no 7 pp 820ndash829 2006

[22] H C Andrews and C L Patterson ldquoSingular value decomposi-tions and digital image processingrdquo IEEE Transactions on SignalProcessing vol 24 no 1 pp 26ndash53 1976

[23] R Kakarala and P O Ogunbona ldquoSignal analysis using a multi-resolution form of the singular value decompositionrdquo IEEETransactions on Image Processing vol 10 no 5 pp 723ndash7352001

[24] B Chen and G WWornell ldquoQuantization index modulation aclass of provably goodmethods for digital watermarking and in-formation embeddingrdquo IEEE Transactions on Information The-ory vol 47 no 4 pp 1423ndash1443 2001

[25] M Yaghoobi ldquoA New Approach for Image Encryption UsingChaotic Logistic Maprdquo in Proceedings of the 2008 InternationalConference on Advanced Computer Theory and Engineering(ICACTE) pp 585ndash590 Phuket Thailand December 2008

[26] R Dugad K Ratakonda and N Ahuja ldquoRobust video shotchange detectionrdquo in Proceedings of the 1998 IEEE SecondWork-shop on Multimedia Signal Processing pp 376ndash381 RedondoBeach CA USA 1998

[27] A Anees I Hussain A Algarni and M Aslam ldquoA RobustWatermarking Scheme for Online Multimedia Copyright Pro-tection Using New Chaotic Maprdquo Security and CommunicationNetworks vol 2018 2018

[28] R Thanki V Dwivedi and K Borisagar ldquoA hybrid watermark-ing scheme with CS theory for security of multimedia datardquoJournal of King Saud University - Computer and InformationSciences 2017

[29] httpmediaxiphorgvideoderf Accessed on 06 July 2018[30] B G Haskell and A N Netravali Digital Pictures Representa-

tion Compression and Standards Perseus Publishing 1997[31] httpwwwvirtualduborg Accessed on 06 July 2018[32] F A Petitcolas R J Anderson and M G Kuhn ldquoAttacks on

CopyrightMarking Systemsrdquo in Information Hiding vol 1525 ofLecture Notes in Computer Science pp 218ndash238 Springer BerlinHeidelberg Berlin Heidelberg 1998

[33] httpwwwpetitcolasnetwatermarkingstirmark Accessed on19 September 2018

Security and Communication Networks 15

100

90

80

70

60

50

4010 20 30 40 50 60 70 80 90 100

NC

()

JPEG quality factor

10 20 30 40 50 60 70 80 90 100

JPEG quality factor

ForemanAkiyoFootball

ForemanAkiyoFootball

30

25

20

15

10

5

0

BER

()

Figure 7 NC and BER values of the extracted watermarks under the JPEG compression attack

Table 7 Extracted watermark (with NC and BER) under image processing attacks with StirMark benchmark

Attacks NC BERJPEG 10 05677 01515JPEG 30 09479 00202JPEG 60 08485 00606JPEG 90 10000 000003x3 median filter+ JPEG 09732 00101Gaussian filter +JPEG 07681 01010Sharpening +JPEG 05127 02626removed one row one colon +JPEG 05523 02020

compression attacks It is clear that the proposed schemeachieves a better robustness

44 Complexity of the Proposed Algorithm The proposedwatermarking schema has a low complexity because of thefollowing

(i) The algorithm is done in the MR-SVD domain whichhas an overall complexity O(n) where n indicates thesignal length [23]

(ii) To extract fast motion frames we use a simple thresh-old decision method based on motion energy

(iii) The embedding and extraction are not applied to allthe frames of the video but just to fast motion frames

In fact the execution time of the proposed scheme required inwatermark embedding and extraction processes the followingresults watermark embedding 036 s per frame watermarkextraction 019 s per frame The simulation is conducted on

Intel(R) Core i5 180 GHz processor with 4 GB RAMusingMATLAB verR2013b

5 Conclusion

In this paper we proposed a novel blind video watermarkingscheme in fast motion frames using the SVD the MR-SVD the QIM and the Logistic Map encryption In theproposed technique we solved the problem of embedding thewatermark in all frames by choosing only the frames that havebig motion energy to be the host frames suitable to the HSVUsing this approach we have limited the number of frames tobe processed and also assured a higher imperceptibility of thewatermark In addition the combination of the characteris-tics of the SVD MR-SVD QIM and Logistic Map Encryp-tion makes our scheme secure and robust to a variety ofattacks The experimental results confirm that the proposedwatermarking scheme has good imperceptibility with PSNRgreater than 40 dBTheproposed scheme is robust not only toimage processing attacks but also to frame synchronization

16 Security and Communication Networks

100

95

90

85

80

75

NC

()

NC

()

100

95

90

85

80

75

NC

()

10 15 20 25 30 35 40

frame averaging rate

proposed method[20][16][6]

proposed method

proposed method

[20]

[20]

[16][6]

[6]

[18]

10 20 30 40 50 60 70

frame dropping rate

70

65

100

95

90

85

80

75

70

65

60

JPEG Quality factor20 30 40 50 60 70 80 90

Figure 8 Comparison of experimental results between the proposed technique and other schemes under frame averaging attack framedropping attack and JPEG compression attack

and video compression attacks The comparison results withother algorithms related to video watermarking indicate thesuperiority of our scheme

Data Availability

The data used to support the findings of this study are avail-able from the corresponding author upon request

Conflicts of Interest

The authors declare no conflicts of interest

References

[1] J C IngemarM LMiller A B Jeffrey J Fridrich andT KalkerDigital Watermarking and Steganography Morgan KaufmannPublishers Inc 2008

[2] S Katzenbeisser and F Petitcolas ldquoInformation Hiding Tech-niques for Steganography and Digital Watermarkingrdquo ArtechHouse 2000

[3] H Tao L Chongmin J M Zain and A N Abdalla ldquoRobustimagewatermarking theories and techniques A reviewrdquo Journalof Applied Research and Technology vol 12 no 1 pp 122ndash1382014

Security and Communication Networks 17

[4] M Kutter F Jordan and F Bossen ldquoDigital watermarking ofcolor images using amplitude modulationrdquo Journal of ElectronicImaging vol 7 no 2 pp 326ndash332 1998

[5] A Karmakar A Phadikar B S Phadikar and G K Maity ldquoAblind video watermarking scheme resistant to rotation and col-lusion attacksrdquo Journal of King Saud University - Computer andInformation Sciences vol 28 no 2 pp 199ndash210 2016

[6] T R Singh K M Singh and S Roy ldquoVideo watermarkingscheme based on visual cryptography and scene change detec-tionrdquo AEU - International Journal of Electronics and Communi-cations vol 67 no 8 pp 645ndash651 2013

[7] W Kong B Yang D Wu and X Niu ldquoSVD Based Blind VideoWatermarking Algorithmrdquo in Proceedings of the First Interna-tional Conference on Innovative Computing Information andControl - Volume I (ICICICrsquo06) pp 265ndash268 Beijing ChinaSeptember 2006

[8] R Liu and T Tan ldquoAn SVD-based watermarking scheme forprotecting rightful ownershiprdquo IEEE Transactions on Multime-dia vol 4 no 1 pp 121ndash128 2002

[9] O S Faragallah ldquoEfficient video watermarking based on sin-gular value decomposition in the discrete wavelet transformdomainrdquo AEU - International Journal of Electronics and Com-munications vol 67 no 3 pp 189ndash196 2013

[10] K Niu X Yang and L Xiang ldquoHybrid quasi-3D DWTDCTand SVD video watermarkingrdquo in Proceedings of the 2010 IEEEInternational Conference on Software Engineering and ServiceSciences ICSESS 2010 pp 588ndash591 China July 2010

[11] N I Yassin NM Salem andM I El Adawy ldquoQIM blind videowatermarking scheme based on Wavelet transform and prin-cipal component analysisrdquo Alexandria Engineering Journal vol53 no 4 pp 833ndash842 2014

[12] D K Thind and S Jindal ldquoA semi blind DWT-SVD videowatermarkingrdquo inProceedings of the International Conference onInformation and Communication Technologies ICICT 2014 pp1661ndash1667 India December 2014

[13] N Leelavathy E V Prasad and S Kumar ldquoVideo Watermark-ing Techniques A Reviewrdquo International Journal of ComputerApplications vol 104 no 7 pp 24ndash30 2014

[14] F Hartung and M Kutter ldquoMultimedia watermarking techni-quesrdquo Proceedings of the IEEE vol 87 no 7 pp 1079ndash1107 1999

[15] P Rasti S Samiei M Agoyi S Escalera and G AnbarjafarildquoRobust non-blind color video watermarking using QR decom-position and entropy analysisrdquo Journal of Visual Communica-tion and Image Representation vol 38 pp 838ndash847 2016

[16] S M Youssef A A ElFarag and N M Ghatwary ldquoAdaptivevideo watermarking integrating a fuzzy wavelet-based humanvisual system perceptual modelrdquo Multimedia Tools and Appli-cations vol 73 no 3 pp 1545ndash1573 2014

[17] T Tabassum and S M Islam ldquoA digital video watermarkingtechnique based on identical frame extraction in 3-LevelDWTrdquoin Proceedings of the 2012 15th International Conference onComputer and Information Technology (ICCIT) pp 101ndash106Chittagong Bangladesh December 2012

[18] L Agilandeeswari and K Ganesan ldquoA robust color video water-marking scheme based on hybrid embedding techniquesrdquoMul-timedia Tools and Applications vol 75 no 14 pp 8745ndash87802016

[19] X Jiang Q Liu and Q Wu ldquoA new video watermarking algo-rithm based on shot segmentation and block classificationrdquoMultimedia Tools and Applications vol 62 no 3 pp 545ndash5602013

[20] K Chetan and K Raghavendra ldquoDWT Based Blind DigitalVideo Watermarking Scheme for Video Authenticationrdquo Inter-national Journal of Computer Applications vol 4 no 10 pp 19ndash26 2010

[21] Y JiaW Lin and A A Kassim ldquoEstimating just-noticeable dis-tortion for videordquo IEEE Transactions on Circuits and Systems forVideo Technology vol 16 no 7 pp 820ndash829 2006

[22] H C Andrews and C L Patterson ldquoSingular value decomposi-tions and digital image processingrdquo IEEE Transactions on SignalProcessing vol 24 no 1 pp 26ndash53 1976

[23] R Kakarala and P O Ogunbona ldquoSignal analysis using a multi-resolution form of the singular value decompositionrdquo IEEETransactions on Image Processing vol 10 no 5 pp 723ndash7352001

[24] B Chen and G WWornell ldquoQuantization index modulation aclass of provably goodmethods for digital watermarking and in-formation embeddingrdquo IEEE Transactions on Information The-ory vol 47 no 4 pp 1423ndash1443 2001

[25] M Yaghoobi ldquoA New Approach for Image Encryption UsingChaotic Logistic Maprdquo in Proceedings of the 2008 InternationalConference on Advanced Computer Theory and Engineering(ICACTE) pp 585ndash590 Phuket Thailand December 2008

[26] R Dugad K Ratakonda and N Ahuja ldquoRobust video shotchange detectionrdquo in Proceedings of the 1998 IEEE SecondWork-shop on Multimedia Signal Processing pp 376ndash381 RedondoBeach CA USA 1998

[27] A Anees I Hussain A Algarni and M Aslam ldquoA RobustWatermarking Scheme for Online Multimedia Copyright Pro-tection Using New Chaotic Maprdquo Security and CommunicationNetworks vol 2018 2018

[28] R Thanki V Dwivedi and K Borisagar ldquoA hybrid watermark-ing scheme with CS theory for security of multimedia datardquoJournal of King Saud University - Computer and InformationSciences 2017

[29] httpmediaxiphorgvideoderf Accessed on 06 July 2018[30] B G Haskell and A N Netravali Digital Pictures Representa-

tion Compression and Standards Perseus Publishing 1997[31] httpwwwvirtualduborg Accessed on 06 July 2018[32] F A Petitcolas R J Anderson and M G Kuhn ldquoAttacks on

CopyrightMarking Systemsrdquo in Information Hiding vol 1525 ofLecture Notes in Computer Science pp 218ndash238 Springer BerlinHeidelberg Berlin Heidelberg 1998

[33] httpwwwpetitcolasnetwatermarkingstirmark Accessed on19 September 2018

16 Security and Communication Networks

100

95

90

85

80

75

NC

()

NC

()

100

95

90

85

80

75

NC

()

10 15 20 25 30 35 40

frame averaging rate

proposed method[20][16][6]

proposed method

proposed method

[20]

[20]

[16][6]

[6]

[18]

10 20 30 40 50 60 70

frame dropping rate

70

65

100

95

90

85

80

75

70

65

60

JPEG Quality factor20 30 40 50 60 70 80 90

Figure 8 Comparison of experimental results between the proposed technique and other schemes under frame averaging attack framedropping attack and JPEG compression attack

and video compression attacks The comparison results withother algorithms related to video watermarking indicate thesuperiority of our scheme

Data Availability

The data used to support the findings of this study are avail-able from the corresponding author upon request

Conflicts of Interest

The authors declare no conflicts of interest

References

[1] J C IngemarM LMiller A B Jeffrey J Fridrich andT KalkerDigital Watermarking and Steganography Morgan KaufmannPublishers Inc 2008

[2] S Katzenbeisser and F Petitcolas ldquoInformation Hiding Tech-niques for Steganography and Digital Watermarkingrdquo ArtechHouse 2000

[3] H Tao L Chongmin J M Zain and A N Abdalla ldquoRobustimagewatermarking theories and techniques A reviewrdquo Journalof Applied Research and Technology vol 12 no 1 pp 122ndash1382014

Security and Communication Networks 17

[4] M Kutter F Jordan and F Bossen ldquoDigital watermarking ofcolor images using amplitude modulationrdquo Journal of ElectronicImaging vol 7 no 2 pp 326ndash332 1998

[5] A Karmakar A Phadikar B S Phadikar and G K Maity ldquoAblind video watermarking scheme resistant to rotation and col-lusion attacksrdquo Journal of King Saud University - Computer andInformation Sciences vol 28 no 2 pp 199ndash210 2016

[6] T R Singh K M Singh and S Roy ldquoVideo watermarkingscheme based on visual cryptography and scene change detec-tionrdquo AEU - International Journal of Electronics and Communi-cations vol 67 no 8 pp 645ndash651 2013

[7] W Kong B Yang D Wu and X Niu ldquoSVD Based Blind VideoWatermarking Algorithmrdquo in Proceedings of the First Interna-tional Conference on Innovative Computing Information andControl - Volume I (ICICICrsquo06) pp 265ndash268 Beijing ChinaSeptember 2006

[8] R Liu and T Tan ldquoAn SVD-based watermarking scheme forprotecting rightful ownershiprdquo IEEE Transactions on Multime-dia vol 4 no 1 pp 121ndash128 2002

[9] O S Faragallah ldquoEfficient video watermarking based on sin-gular value decomposition in the discrete wavelet transformdomainrdquo AEU - International Journal of Electronics and Com-munications vol 67 no 3 pp 189ndash196 2013

[10] K Niu X Yang and L Xiang ldquoHybrid quasi-3D DWTDCTand SVD video watermarkingrdquo in Proceedings of the 2010 IEEEInternational Conference on Software Engineering and ServiceSciences ICSESS 2010 pp 588ndash591 China July 2010

[11] N I Yassin NM Salem andM I El Adawy ldquoQIM blind videowatermarking scheme based on Wavelet transform and prin-cipal component analysisrdquo Alexandria Engineering Journal vol53 no 4 pp 833ndash842 2014

[12] D K Thind and S Jindal ldquoA semi blind DWT-SVD videowatermarkingrdquo inProceedings of the International Conference onInformation and Communication Technologies ICICT 2014 pp1661ndash1667 India December 2014

[13] N Leelavathy E V Prasad and S Kumar ldquoVideo Watermark-ing Techniques A Reviewrdquo International Journal of ComputerApplications vol 104 no 7 pp 24ndash30 2014

[14] F Hartung and M Kutter ldquoMultimedia watermarking techni-quesrdquo Proceedings of the IEEE vol 87 no 7 pp 1079ndash1107 1999

[15] P Rasti S Samiei M Agoyi S Escalera and G AnbarjafarildquoRobust non-blind color video watermarking using QR decom-position and entropy analysisrdquo Journal of Visual Communica-tion and Image Representation vol 38 pp 838ndash847 2016

[16] S M Youssef A A ElFarag and N M Ghatwary ldquoAdaptivevideo watermarking integrating a fuzzy wavelet-based humanvisual system perceptual modelrdquo Multimedia Tools and Appli-cations vol 73 no 3 pp 1545ndash1573 2014

[17] T Tabassum and S M Islam ldquoA digital video watermarkingtechnique based on identical frame extraction in 3-LevelDWTrdquoin Proceedings of the 2012 15th International Conference onComputer and Information Technology (ICCIT) pp 101ndash106Chittagong Bangladesh December 2012

[18] L Agilandeeswari and K Ganesan ldquoA robust color video water-marking scheme based on hybrid embedding techniquesrdquoMul-timedia Tools and Applications vol 75 no 14 pp 8745ndash87802016

[19] X Jiang Q Liu and Q Wu ldquoA new video watermarking algo-rithm based on shot segmentation and block classificationrdquoMultimedia Tools and Applications vol 62 no 3 pp 545ndash5602013

[20] K Chetan and K Raghavendra ldquoDWT Based Blind DigitalVideo Watermarking Scheme for Video Authenticationrdquo Inter-national Journal of Computer Applications vol 4 no 10 pp 19ndash26 2010

[21] Y JiaW Lin and A A Kassim ldquoEstimating just-noticeable dis-tortion for videordquo IEEE Transactions on Circuits and Systems forVideo Technology vol 16 no 7 pp 820ndash829 2006

[22] H C Andrews and C L Patterson ldquoSingular value decomposi-tions and digital image processingrdquo IEEE Transactions on SignalProcessing vol 24 no 1 pp 26ndash53 1976

[23] R Kakarala and P O Ogunbona ldquoSignal analysis using a multi-resolution form of the singular value decompositionrdquo IEEETransactions on Image Processing vol 10 no 5 pp 723ndash7352001

[24] B Chen and G WWornell ldquoQuantization index modulation aclass of provably goodmethods for digital watermarking and in-formation embeddingrdquo IEEE Transactions on Information The-ory vol 47 no 4 pp 1423ndash1443 2001

[25] M Yaghoobi ldquoA New Approach for Image Encryption UsingChaotic Logistic Maprdquo in Proceedings of the 2008 InternationalConference on Advanced Computer Theory and Engineering(ICACTE) pp 585ndash590 Phuket Thailand December 2008

[26] R Dugad K Ratakonda and N Ahuja ldquoRobust video shotchange detectionrdquo in Proceedings of the 1998 IEEE SecondWork-shop on Multimedia Signal Processing pp 376ndash381 RedondoBeach CA USA 1998

[27] A Anees I Hussain A Algarni and M Aslam ldquoA RobustWatermarking Scheme for Online Multimedia Copyright Pro-tection Using New Chaotic Maprdquo Security and CommunicationNetworks vol 2018 2018

[28] R Thanki V Dwivedi and K Borisagar ldquoA hybrid watermark-ing scheme with CS theory for security of multimedia datardquoJournal of King Saud University - Computer and InformationSciences 2017

[29] httpmediaxiphorgvideoderf Accessed on 06 July 2018[30] B G Haskell and A N Netravali Digital Pictures Representa-

tion Compression and Standards Perseus Publishing 1997[31] httpwwwvirtualduborg Accessed on 06 July 2018[32] F A Petitcolas R J Anderson and M G Kuhn ldquoAttacks on

CopyrightMarking Systemsrdquo in Information Hiding vol 1525 ofLecture Notes in Computer Science pp 218ndash238 Springer BerlinHeidelberg Berlin Heidelberg 1998

[33] httpwwwpetitcolasnetwatermarkingstirmark Accessed on19 September 2018

Security and Communication Networks 17

[4] M Kutter F Jordan and F Bossen ldquoDigital watermarking ofcolor images using amplitude modulationrdquo Journal of ElectronicImaging vol 7 no 2 pp 326ndash332 1998

[5] A Karmakar A Phadikar B S Phadikar and G K Maity ldquoAblind video watermarking scheme resistant to rotation and col-lusion attacksrdquo Journal of King Saud University - Computer andInformation Sciences vol 28 no 2 pp 199ndash210 2016

[6] T R Singh K M Singh and S Roy ldquoVideo watermarkingscheme based on visual cryptography and scene change detec-tionrdquo AEU - International Journal of Electronics and Communi-cations vol 67 no 8 pp 645ndash651 2013

[7] W Kong B Yang D Wu and X Niu ldquoSVD Based Blind VideoWatermarking Algorithmrdquo in Proceedings of the First Interna-tional Conference on Innovative Computing Information andControl - Volume I (ICICICrsquo06) pp 265ndash268 Beijing ChinaSeptember 2006

[8] R Liu and T Tan ldquoAn SVD-based watermarking scheme forprotecting rightful ownershiprdquo IEEE Transactions on Multime-dia vol 4 no 1 pp 121ndash128 2002

[9] O S Faragallah ldquoEfficient video watermarking based on sin-gular value decomposition in the discrete wavelet transformdomainrdquo AEU - International Journal of Electronics and Com-munications vol 67 no 3 pp 189ndash196 2013

[10] K Niu X Yang and L Xiang ldquoHybrid quasi-3D DWTDCTand SVD video watermarkingrdquo in Proceedings of the 2010 IEEEInternational Conference on Software Engineering and ServiceSciences ICSESS 2010 pp 588ndash591 China July 2010

[11] N I Yassin NM Salem andM I El Adawy ldquoQIM blind videowatermarking scheme based on Wavelet transform and prin-cipal component analysisrdquo Alexandria Engineering Journal vol53 no 4 pp 833ndash842 2014

[12] D K Thind and S Jindal ldquoA semi blind DWT-SVD videowatermarkingrdquo inProceedings of the International Conference onInformation and Communication Technologies ICICT 2014 pp1661ndash1667 India December 2014

[13] N Leelavathy E V Prasad and S Kumar ldquoVideo Watermark-ing Techniques A Reviewrdquo International Journal of ComputerApplications vol 104 no 7 pp 24ndash30 2014

[14] F Hartung and M Kutter ldquoMultimedia watermarking techni-quesrdquo Proceedings of the IEEE vol 87 no 7 pp 1079ndash1107 1999

[15] P Rasti S Samiei M Agoyi S Escalera and G AnbarjafarildquoRobust non-blind color video watermarking using QR decom-position and entropy analysisrdquo Journal of Visual Communica-tion and Image Representation vol 38 pp 838ndash847 2016

[16] S M Youssef A A ElFarag and N M Ghatwary ldquoAdaptivevideo watermarking integrating a fuzzy wavelet-based humanvisual system perceptual modelrdquo Multimedia Tools and Appli-cations vol 73 no 3 pp 1545ndash1573 2014

[17] T Tabassum and S M Islam ldquoA digital video watermarkingtechnique based on identical frame extraction in 3-LevelDWTrdquoin Proceedings of the 2012 15th International Conference onComputer and Information Technology (ICCIT) pp 101ndash106Chittagong Bangladesh December 2012

[18] L Agilandeeswari and K Ganesan ldquoA robust color video water-marking scheme based on hybrid embedding techniquesrdquoMul-timedia Tools and Applications vol 75 no 14 pp 8745ndash87802016

[19] X Jiang Q Liu and Q Wu ldquoA new video watermarking algo-rithm based on shot segmentation and block classificationrdquoMultimedia Tools and Applications vol 62 no 3 pp 545ndash5602013

[20] K Chetan and K Raghavendra ldquoDWT Based Blind DigitalVideo Watermarking Scheme for Video Authenticationrdquo Inter-national Journal of Computer Applications vol 4 no 10 pp 19ndash26 2010

[21] Y JiaW Lin and A A Kassim ldquoEstimating just-noticeable dis-tortion for videordquo IEEE Transactions on Circuits and Systems forVideo Technology vol 16 no 7 pp 820ndash829 2006

[22] H C Andrews and C L Patterson ldquoSingular value decomposi-tions and digital image processingrdquo IEEE Transactions on SignalProcessing vol 24 no 1 pp 26ndash53 1976

[23] R Kakarala and P O Ogunbona ldquoSignal analysis using a multi-resolution form of the singular value decompositionrdquo IEEETransactions on Image Processing vol 10 no 5 pp 723ndash7352001

[24] B Chen and G WWornell ldquoQuantization index modulation aclass of provably goodmethods for digital watermarking and in-formation embeddingrdquo IEEE Transactions on Information The-ory vol 47 no 4 pp 1423ndash1443 2001

[25] M Yaghoobi ldquoA New Approach for Image Encryption UsingChaotic Logistic Maprdquo in Proceedings of the 2008 InternationalConference on Advanced Computer Theory and Engineering(ICACTE) pp 585ndash590 Phuket Thailand December 2008

[26] R Dugad K Ratakonda and N Ahuja ldquoRobust video shotchange detectionrdquo in Proceedings of the 1998 IEEE SecondWork-shop on Multimedia Signal Processing pp 376ndash381 RedondoBeach CA USA 1998

[27] A Anees I Hussain A Algarni and M Aslam ldquoA RobustWatermarking Scheme for Online Multimedia Copyright Pro-tection Using New Chaotic Maprdquo Security and CommunicationNetworks vol 2018 2018

[28] R Thanki V Dwivedi and K Borisagar ldquoA hybrid watermark-ing scheme with CS theory for security of multimedia datardquoJournal of King Saud University - Computer and InformationSciences 2017

[29] httpmediaxiphorgvideoderf Accessed on 06 July 2018[30] B G Haskell and A N Netravali Digital Pictures Representa-

tion Compression and Standards Perseus Publishing 1997[31] httpwwwvirtualduborg Accessed on 06 July 2018[32] F A Petitcolas R J Anderson and M G Kuhn ldquoAttacks on

CopyrightMarking Systemsrdquo in Information Hiding vol 1525 ofLecture Notes in Computer Science pp 218ndash238 Springer BerlinHeidelberg Berlin Heidelberg 1998

[33] httpwwwpetitcolasnetwatermarkingstirmark Accessed on19 September 2018

International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Active and Passive Electronic Components

VLSI Design

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Shock and Vibration

Hindawiwwwhindawicom Volume 2018

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom

Journal ofEngineeringVolume 2018

SensorsJournal of

Hindawiwwwhindawicom Volume 2018

International Journal of

RotatingMachinery

Hindawiwwwhindawicom Volume 2018

Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Navigation and Observation

International Journal of

Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom