cloud basedmultimedia
TRANSCRIPT
-
8/19/2019 Cloud Basedmultimedia
1/7
CLOUD-BASED MULTIMEDIA CONTENT PROTECTION
SYSTEM
Abstract — We propose a new design for large-scale multimedia content protection
systems. Our design leverages cloud infrastructures to provide cost efficiency,
rapid deployment, scalability, and elasticity to accommodate varying workloads.
The proposed system can be used to protect different multimedia content types,
including 2-D videos, -D videos, images, audio clips, songs, and music clips. The
system can be deployed on private and!or public clouds. Our system has two novel
components" #i$ method to create signatures of -D videos, and #ii$ distributed
matching engine for multimedia ob%ects. The signature method creates robust and
representative signatures of -D videos that capture the depth signals in these
videos and it is computationally efficient to compute and compare as well as it
re&uires small storage. The distributed matching engine achieves high scalability
and it is designed to support different multimedia ob%ects. We implemented the
proposed system and deployed it on two clouds" 'ma(on cloud and our private
cloud. Our e)periments with more than **,+++ -D videos and * million images
show the high accuracy and scalability of the proposed system. n addition, we
compared our system to the protection system used by ouTube and our results
show that the ouTube protection system fails to detect most copies of -D videos,
while our system detects more than /0 of them. This comparison shows the need
for the proposed -D signature method, since the state-of-the-art commercial
system was not able to handle -D videos
-
8/19/2019 Cloud Basedmultimedia
2/7
EXISTING SYSTEM:
The problem of protecting various types of multimedia content has attracted
significant attention from academia and industry. One approach to this problem is
using watermarking, in which some distinctive information is embedded in the
content itself and a method is used to search for this information in order to verify
the authenticity of the content. Watermarking re&uires inserting watermarks in the
multimedia ob%ects before releasing them as well as mechanisms!systems to find
ob%ects and verify the e)istence of correct watermarks in them. Thus, this approach
may not be suitable for already-released content without watermarks in them. The
watermarking approach is more suitable for the somewhat controlled
environments, such as distribution of multimedia content on D1Ds or using special
sites and custom players. Watermarking may not be effective for the rapidly
increasing online videos, especially those uploaded to sites such as ouTube and
played back by any video player. Watermarking is not the focus of this paper. The
focus of this paper is on the other approach for protecting multimedia content,
which is content-based copy detection #3D$. n this approach, signatures #or fingerprints$ are e)tracted from original ob%ects. 4ignatures are also created from
&uery #suspected$ ob%ects downloaded from online sites. Then, the similarity is
computed between original and suspected ob%ects to find potential copies. 5any
previous works proposed different methods for creating and matching signatures.
These methods can be classified into four categories" spatial, temporal, color, and
transform-domain. 4patial signatures #particularly the block-based$ are the most
widely used. 6owever, their weakness is the lack of resilience against large
geometric transformations. Temporal and color signatures are less robust and can
be used to enhance spatial signatures. Transform-domain signatures are
-
8/19/2019 Cloud Basedmultimedia
3/7
computationally intensive and not widely used in practice. 7or more details, see
surveys for audio fingerprinting and 2-D video fingerprinting.
PROPOSED SYSTEM:
The proposed system is fairly comple) with multiple components, including" #i$
crawler to download thousands of multimedia ob%ects from online hosting sites, #ii$
signature method to create representative fingerprints from multimedia ob%ects, and
#iii$ distributed matching engine to store signatures of original ob%ects and match
them against &uery ob%ects. We propose novel methods for the second and third
components, and we utili(e off-the-shelf tools for the crawler. We have developed
a complete running system of all components and tested it with more than **,+++
-D videos and * million images. We deployed parts of the system on the 'ma(on
cloud with varying number of machines #from eight to *2/$, and the other parts of
the system were deployed on our private cloud. This deployment model was used
to show the fle)ibility of our system, which enables it to efficiently utili(e varying
computing resources and minimi(e the cost, since cloud providers offer different
pricing models for computing and network resources. Through e)tensive
e)periments with real deployment, we show the high accuracy #in terms of
precision and recall$ as well as the scalability and elasticity of the proposed
system.
-
8/19/2019 Cloud Basedmultimedia
4/7
CONCLUSION AND FUTURE WORK
Distributing copyrighted multimedia ob%ects by uploading them to online hosting
sites such as ouTube can result in significant loss of revenues for content creators.
4ystems needed to find illegal copies of multimedia ob%ects are comple) and large
scale. n this paper, we presented a new design for multimedia content protection
systems using multi-cloud infrastructures. The proposed system supports different
multimedia content types and it can be deployed on private and!or public clouds.
Two key components of the proposed system are presented. The first one is a new
method for creating signatures of -D videos. Our method constructs coarse-
grained disparity maps using stereo correspondence for a sparse set of points in the
image. Thus, it captures the depth signal of the -D video, without e)plicitly
computing the e)act depth map, which is computationally e)pensive. Our
e)periments showed that the proposed -D signature produces high accuracy in
terms of both precision and recall and it is robust to many video transformations
including new ones that are specific to -D videos such as synthesi(ing new views.
The second key component in our system is the distributed inde), which is used to
match multimedia ob%ects characteri(ed by high dimensions. The distributed inde)
is implemented using the 5ap8educe framework and our e)periments showed that
it can elastically utili(e varying amount of computing resources and it produces
high accuracy. The e)periments also showed that it outperforms the closest system
in the literature in terms of accuracy and computational efficiency. n addition, we
evaluated the whole content protection system with more than **,+++ -D videos
and the results showed the scalability and accuracy of the proposed system.
7inally, we compared our system against the ontent D system used by ouTube.
Our results showed that" #i$ there is a need for designing robust signatures for -D
videos since the current system used by the leading company in the industry fails
-
8/19/2019 Cloud Basedmultimedia
5/7
to detect most modified -D copies, and #ii$ our proposed -D signature method
can fill this gap, because it is robust to many 2-D and -D video transformations.
The work in this paper can be e)tended in multiple directions. 7or e)ample, our
current system is optimi(ed for batch processing. Thus, it may not be suitable for
online detection of illegally distributed multimedia streams of live events such as
soccer games. n live events, only small segments of the video are available and
immediate detection of copyright infringement is crucial to minimi(e financial
losses. To support online detection, the matching engine of our system needs to be
implemented using a distributed programming framework that supports online
processing, such as 4park. n addition, composite signature schemes that combine
multiple modalities may be needed to &uickly identify short video segments.
7urthermore, the crawler component needs to be customi(ed to find online sites
that offer pirated video streams and obtain segments of these streams for checking
against reference streams, for which the signatures would also need to be generated
online. 'nother future direction for the work in this paper is to design signatures
for recent and comple) formats of -D videos such as multiview plus depth. '
multiview plus depth video has multiple te)ture and depth components, which
allow users to view a scene from different angles. 4ignatures for such videos would
need to capture this comple)ity, while being efficient to compute, compare, and
store.
REFERENCES
9*: '. 'bdelsadek, ;Distributed inde) for matching multimedia ob%ects,< 5.4.
thesis, 4chool of omput. 4ci., 4imon 7raser =niv., 3urnaby, 3, anada, 2+*>.
92: '. 'bdelsadek and 5. 6efeeda, ;Dimo" Distributed inde) for matching
multimedia ob%ects using 5ap8educe,< in Proc. ACMMultimedia Syst. Conf.
(MMSys’14), 4ingapore, 5ar. 2+*>, pp. **?@*2?.
-
8/19/2019 Cloud Basedmultimedia
6/7
9: 5. 'ly, 5. 5unich, and A. Aerona, ;Distributed Bd-Trees for retrieval from
very large image collections,< in Proc. Brit. Mach. is. Conf. (BMC), Dundee,
=.B., 'ug. 2+**.
9>: C. 3entley, ;5ultidimensional binary search trees used for associative
searching,< in Commun. ACM , 4ep. *?, vol. */, no. , pp. ?+@?*.
9?: A. ano, E. 3atle, T. Balker, and C. 6aitsma, ;' review of algorithms for audio
fingerprinting,< in Proc. !""" #or$sho% Multimedia Si&nal Process., Dec. 2++2,
pp. *F@*.
9F: C. Dean and 4. Ghemawat, ;5ap8educe" 4implified data processing on large
clusters,< in Proc. Sym%. '%er. Syst. esi&n !m%lementation ('S!’4), 4an
7rancisco, ', =4', Dec. 2++>, pp. *@*?+.
9: C. Deng, W. Dong, 8. 4ocher, H. Hi, B. Hi, and H. 7ei-7ei, ;magenet" ' large-
scale hierarchical image database,< in Proc. !""" Conf. Com%ut. is. Pattern
*eco&. (CP*’+), 5iami, 7H, =4', Cun. 2++, pp. 2>/@2??.
9/: '. 6ampapur, B. 6yun, and 8. 3olle, ;omparison of se&uence matching
techni&ues for video copy detection,< in Proc. SP!" Conf. Stora&e *etrie,al Media
atabases (SP!"’-), 4an Cose, ', =4', Can. 2++2, pp. *>@2+*.
9: 4. offe, ;7ull-length video fingerprinting. Google nc.,< =.4. Aatent /222*,
Cul. 2>, 2+*2.
9*+: '. Bahng, C. Hach, W. 5angione-4mith, 4. 5antik, . 5arkov, 5. Aotkon%ak,
A. Tucker, 6. Wang, and G. Wolfe, ;Watermarking techni&ues for intellectual
property protection,< in Proc. /th Annu. esi&n Autom. Conf. (AC’+0), 4an
7rancisco, ', =4', Cun. */, pp. F@/*.
9**: I. Bhodabakhshi and 5. 6efeeda, ;4pider" ' system for finding D video
copies,< in ACM rans. Multimedia Com%ut.2 Commun.2 A%%l. ('MM), 7eb.
2+*, vol. , no. *, pp. "*@"2+.
-
8/19/2019 Cloud Basedmultimedia
7/7
9*2: 4. Hee and . oo, ;8obust video fingerprinting for content-based video
identification,