an image-based approach to video copy detection with spatio -temporal post-filtering

19
An Image-Based Approach to Video Copy Detection With Spatio-Temporal Post-Filtering Matthijs Douze, Hervé Jégou, and Cordelia Schmid, Senior Member, IEEE

Upload: sagira

Post on 16-Feb-2016

51 views

Category:

Documents


0 download

DESCRIPTION

An Image-Based Approach to Video Copy Detection With Spatio -Temporal Post-Filtering. Matthijs Douze , Hervé Jégou , and Cordelia Schmid , Senior Member, IEEE. INTRODUCTION. Common distortions are 1. scaling 2. compression 3. cropping 4. camcording. FRAME INDEXING (step1~6). - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: An Image-Based Approach to Video Copy Detection With  Spatio -Temporal Post-Filtering

An Image-Based Approach to Video Copy Detection With

Spatio-Temporal Post-Filtering

Matthijs Douze, Hervé Jégou, and Cordelia Schmid, Senior

Member, IEEE

Page 2: An Image-Based Approach to Video Copy Detection With  Spatio -Temporal Post-Filtering

INTRODUCTION

Common distortions are 1. scaling2. compression 3. cropping4. camcording

Page 3: An Image-Based Approach to Video Copy Detection With  Spatio -Temporal Post-Filtering
Page 4: An Image-Based Approach to Video Copy Detection With  Spatio -Temporal Post-Filtering

FRAME INDEXING (step1~6)

a. Frame Sampling

1. Uniform sampling

2. Keyframes

Page 5: An Image-Based Approach to Video Copy Detection With  Spatio -Temporal Post-Filtering

b. Local Features (salient interest

points)

invariant :

1. Scale change

2. Image rotation

3. Noise

c. Bag-of-Features and Hamming

Embedding

Page 6: An Image-Based Approach to Video Copy Detection With  Spatio -Temporal Post-Filtering
Page 7: An Image-Based Approach to Video Copy Detection With  Spatio -Temporal Post-Filtering
Page 8: An Image-Based Approach to Video Copy Detection With  Spatio -Temporal Post-Filtering
Page 9: An Image-Based Approach to Video Copy Detection With  Spatio -Temporal Post-Filtering

SPATIO-TEMPORAL VERIFICATION

A. Spatio-Temporal Transformation

B. Temporal GroupingC. Spatial Verification(next)D.Score Aggregation

Strategy

Page 10: An Image-Based Approach to Video Copy Detection With  Spatio -Temporal Post-Filtering

Spatial Verification 1. take all point matches from the matching frames. 2. estimate possible similarity transformations from all matching points with a Hough transform.(next)

Page 11: An Image-Based Approach to Video Copy Detection With  Spatio -Temporal Post-Filtering
Page 12: An Image-Based Approach to Video Copy Detection With  Spatio -Temporal Post-Filtering
Page 13: An Image-Based Approach to Video Copy Detection With  Spatio -Temporal Post-Filtering

3. compute and score possible affine transformations. 4. select the maximum score over all possible hypotheses.

Page 14: An Image-Based Approach to Video Copy Detection With  Spatio -Temporal Post-Filtering

ExperimentA. Parameter Optimization(next)B. Handling of Trecvid AttacksC. Trecvid Copy Detection Results

Page 15: An Image-Based Approach to Video Copy Detection With  Spatio -Temporal Post-Filtering
Page 16: An Image-Based Approach to Video Copy Detection With  Spatio -Temporal Post-Filtering
Page 17: An Image-Based Approach to Video Copy Detection With  Spatio -Temporal Post-Filtering
Page 18: An Image-Based Approach to Video Copy Detection With  Spatio -Temporal Post-Filtering
Page 19: An Image-Based Approach to Video Copy Detection With  Spatio -Temporal Post-Filtering

Conclusion

Our video copy detection system outperforms other submitted results on all transformations. This is due to a very accurate image-level matching. Run KEYSADVES, which is more scalable, shows that our system still obtains excellent results with a memory footprint and query time reduced 20 times.