surf: speeded-up robust features computer vision and image understanding ( cviu ) 2008. advisor :...

29
SURF: SPEEDED-UP ROBUST FEATURES Computer Vision and Image Understanding (CVIU) 2008. Advisor : Sheng-Jyh Wang Student : 劉劉劉 2011/10/17

Post on 19-Dec-2015

216 views

Category:

Documents


0 download

TRANSCRIPT

SURF: SPEEDED-UP ROBUST FEATURES

Computer Vision and Image Understanding (CVIU) 2008.

Advisor : Sheng-Jyh Wang

Student : 劉彥廷

2011/10/17

OUTLINE

Introduction

Related Works

Speed-Up Robust Features

• Detection• Description

Experiments

Conclusion

2

OUTLINE

Introduction

Related Works

Speed-Up Robust Features

• Detection• Description

Experiments

Conclusion

3

• Why do we care about feature matching?

Object Recognition

Wide baseline matching

Tracking

4

Introduction

Types of variance

• Illumination• Scale• Rotation• Affine• Perspective

We want to find

Repeatability、 Distinctiveness

features

5

Challenges

OUTLINE

Introduction

Related Works

Speed-Up Robust Features

• Detection• Description

Experiments

Conclusion

6

Related Works

• Harris Corner Detector - Harris 1988• Laplacian of Gaussian - Lindeberg 1998• Difference of Gaussian - Lowe 2004

7

Related Works

• Harris Corner Detector - Harris 1988• Laplacian of Gaussian - Lindeberg 1998• Difference of Gaussian - Lowe 2004

flat edge corner

Illumination invariance !!! 8

Related Works

• Harris Corner Detector - Harris 1988• Laplacian of Gaussian - Lindeberg 1998• Difference of Gaussian - Lowe 2004

* =

characteristic scale

9

LoG can detect blob-like structures at locations

“Feature Detection with Automatic Scale Selection”, IJCV ‘98

Related Works

• Harris Corner Detector - Harris 1988• Laplacian of Gaussian - Lindeberg 1998• Difference of Gaussian - Lowe 2004

2 2( , , ) ( , , ) ( 1) ( , , )G x y k G x y k G x y

10

Computational efficiency !

Keep the same keypoint in all scale !

Compare to 26 neighbors

Motivation

• Lindeberg uses Laplacian of Gaussian, one could obtain scale invariant features.

• Lowe uses difference of Gaussian to approximate Laplacian of Gaussian. (SIFT)

• This paper uses Hessian - Laplacian to approximate Laplacian of Gaussian, to improve calculation speed.

11

OUTLINE

Introduction

Related Works

Speed-Up Robust Features

• Detection• Description

Experiments

Conclusion

12

Detection

Hessian-based interest point localization

• Lxx(x,y,σ) is the Laplacian of Gaussian of the image.

• It is the convolution of the Gaussian second order derivative with the image.

• This paper use Dxx to approximate Lxx.

13

DetectionScale analysis with constant image size

Approximated second order derivatives with box filters.

14

(DoG)

   

Integral Images

Using integral images for major speed upIntegral Image (summed area tables) is an intermediate representation for the image and contains the sum of gray scale pixel values of image.

15They can be evaluated at a very low computational cost using integral images with box filters

16

Keypoint detection

Summary

 

 

 

 

Keypoint description

Keypointmatching

Fourier v.s. Wavelet

17

• Fourier Transform (FT) is not a good tool –

gives no direct information about when an oscillation occurred.

• Wavelets can keep track of time and frequency information.

Fourier basis

Haar basis

interest point

scale = s

r = 6s

Haar dx

dy

Description

Orientation Assignment

x response y response

• The Haar wavelet responses are represented as vectors

18

• Sum all responses within a sliding orientation window covering an angle of 60 degree

• The longest vector is the dominant orientation

Description

• Split the interest region (20s x 20s) up into 4 x 4 square sub-regions.

• Calculate Haar wavelet response dx and dy and weight the response with a Gaussian kernel.

• Sum the response over each sub-region for dx and dy, then sum the absolute value of resp-onse.

19

Matching

20

• Fast indexing through the sign of the Laplacian for the underlying interest point

The sign of trace of the Hessian matrix

Trace = Lxx + Lyy

can do match can do match not match

matching

OUTLINE

Introduction

Related Works

Speed-Up Robust Features

• Detection• Description

Experiments

Conclusion

21

Experiments

22

• Test keypoint repeatability for

(Viewpoint Change), (Lighting Change) and(Zoom and

Rotation)

Experiments

23

• Repeatability score for image sequences

Experiments

• Fix number of keypoints

24

25

Experiments

SIFT

SURF

Leila Mirmohamadsadeghi , “Image Tag Propagation “ ‘10

Image size : 341x341

Running time : 2.411188 seconds

Experiments

26

Image size : 800x600

Running time : 12.028462 seconds

Experiments

27

Conclusion

• SURF is faster than SIFT by 3 times, and has recall precision not worse than SIFT.

• SURF is good at handling image with blurring or

rotation.

• SURF is poor at handling image with viewpoint .

28

29

Reference

• “Speeded-Up Robust Features”, CVIU ‘08 Herbert Bay

• “Distinctive Image Features from Scale-Invariant Features”, IJCV ’04 David G. Lowe

• “A Combined Corner and Edge Detector” ‘88 Chris Harris

• “Feature Detection with Automatic Scale Selection”, IJCV ’98 Lindeberg