1 / 45

“Secret” of Object Detection

“Secret” of Object Detection. Zheng Wu (Summer intern in MSRNE) Sep. 3, 2010 Joint work with Ce Liu (MSRNE) William T. Freeman (MIT) Adam Kalai (MSRNE) Jianxiong Xiao (MIT). person. motorbike. Outline. Introduction Sliding-Window-based Object Detection Window Generation

branxton
Download Presentation

“Secret” of Object Detection

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. “Secret” of Object Detection Zheng Wu (Summer intern in MSRNE) Sep. 3, 2010 Joint work with Ce Liu (MSRNE) William T. Freeman (MIT) Adam Kalai (MSRNE) Jianxiong Xiao (MIT)

  2. person motorbike

  3. Outline Introduction Sliding-Window-based Object Detection Window Generation Feature Extraction Cascade Classifier PASCAL Challenge Conclusion

  4. Related Topic • Image Matching • Image Classification • Object Detection

  5. Object Detection • Single Pattern • Multiple Patterns Viola & Jones Face Detector Felzenszwalb’s Part-based Detector Dalal & Triggs Pedestrian Detector

  6. Outline Introduction Sliding-Window-based Object Detection Window Generation Feature Extraction Cascade Classifier PASCAL Challenge Conclusion

  7. Sliding Windows 7 • Search over: • Location • Scale • Aspect Ratio • Millions of windows!

  8. Sliding Windows • Subsample on grid • - have to set “optimal” step size manually • Fix Aspect Ratio - assume single pattern detection • Fix Scale - assume object’s resolution does not change much between training and test sets. • Search with branch-and-bound method • - have to use special scoring function

  9. Sliding Windows • We propose sliding windows from segmentation Region Segmentation [Felzenszwalb & Huttenlocher, IJCV03] SuperpixelSegmentation [Levinshtein et al, PAMI09] < 100,000 sliding windows / image on PASCAL Dataset

  10. Outline • Introduction • Sliding-Window-based Object Detection • Window Generation • Feature Extraction • Cascade Classifier • PASCAL Challenge • Conclusion

  11. Generic Feature • “Objectness” features [Alexe et al, CVPR10]

  12. Generic Feature • Each type of generic feature is weak, but combination is stronger • Low dimensional feature (=8) • Not suitable for objects with “concave” shape, i.e. table, chair

  13. Generic Feature

  14. Class-specific Feature • Histogram of Orientated Gradients

  15. Class-specific Feature • Dense grid (>10*10) (secret 1) • Overlapping cells • Histogram bin size • High dimensional feature (>1000) • - redundant oroverfitting? • Normalization • No spatial relationship maintained

  16. Outline • Introduction • Sliding-Window-based Object Detection • Window Generation • Feature Extraction • Cascade Classifier • PASCAL Challenge • Conclusion

  17. Cascade Classifier • Same type of classifier with different features • Viola & Jones Face Detector, IJCV01 • Different types of classifier with same features • Harzallah et al, ICCV09 (INRIA) • Vedaldi et al, ICCV09 (Oxford)

  18. Cascade Classifier • Training SVM is slow… • to train 20,000 examples with 4000 dimensions: • >15min for Linear SVM • >3 hours for Nonlinear SVM • Training SVM requires a lot of memory… • design matrix: 20,000*20,000 matrix • Training with Imbalance data • a few hundreds of positive examples • billions of negative examples

  19. Boosted SVM

  20. Boosted SVM • 42327 examples, half for training, half for testing • Training error is 0.05 for all boosted classifiers

  21. Positive Training Set

  22. Negative Training Set Training Sample Pool False Positives Random Samples … … Secret 2 SVM ver. 2 SVM ver. 1

  23. Outline • Introduction • Sliding-Window-based Object Detection • Window Generation • Feature Extraction • Cascade Classifier • PASCAL Challenge • Conclusion

  24. PASCAL Dataset 2009

  25. PASCAL Dataset 2009

  26. PASCAL Dataset 2009

  27. PASCAL Dataset 2009

  28. PASCAL Dataset 2009

  29. PASCAL Dataset 2009

  30. PASCAL Dataset 2009

  31. PASCAL Dataset 2009

  32. PASCAL 2009 (trainval + test)

  33. PASCAL 2009 (train+val)

  34. PASCAL 2009 (train+val,1/image)

  35. PASCAL 2009 (train+val,5/image)

  36. PASCAL 2009 (train+val,10/image)

  37. True Positives - aeroplane

  38. False Positives - aeroplane

  39. True Positives - bicycle

  40. False Positives - bicycle

  41. True Positives - horse

  42. False Positives - horse

  43. Conclusion • Proposing sliding windows without fixing scale or aspect ratio is possible. • Simple feature (saliency, contrast, etc) is only useful for certain object classes. • Histogram-based feature is not sufficient to represent real world object, no matter what learning algorithm is used. • Boosting is helpful to speed up SVM-training and reduce the memory usage. • Digging out “hard” negative examples. • Throwing away “hard” positive examples.

  44. Future Work • It is time to go beyond the histogram-of-X feature • - not every pixel within bounding box is informative • - the appearance of object’s part is more robust • Evolve the classifier • - even PASCAL dataset is too small • - the right decision boundary is still far away… • - Active learning? Online learning?

  45. Thank You !

More Related