1. Image Parsing: Unifying Segmentation and Detection Z. Tu, X. Chen, A.L. Yuille and S-C. Hz ICCV 2003 (Marr Prize) & IJCV 2005 Sanketh Shetty

2. Outline Why Image Parsing? Introduction to Concepts in DDMCMC DDMCMC applied to Image Parsing Combining Discriminative and Generative Models for Parsing Results Comments

3. Image Parsing Image I Parse Structure W Optimize p(W|I)

4. Properties of Parse Structure Dynamic and reconfigurable –Variable number of nodes and node types Defined by a Markov Chain –Data Driven Markov Chain Monte Carlo (earlier work in segmentation, grouping and recognition)

5. Key Concepts Joint model for Segmentation & Recognition –Combine different modules to obtain cues Fully generative explanation for Image generation –Uses Generative and Discriminative Models + DDMCMC framework –Concurrent Top-Down & Bottom-Up Parsing

6. Pattern Classes 62 characters Faces Regions

7. Key Concepts: –Markov Chains –Markov Chain Monte Carlo Metropolis-Hastings [Metropolis 1953, Hastings 1970] Reversible Jump [Green 1995] –Data Driven Markov Chain Monte Carlo MCMC: A Quick Tour

8. Markov Chains Notes: Slides by Zhu, Dellaert and Tu at ICCV 2005

9. Markov Chain Monte Carlo Notes: Slides by Zhu, Dellaert and Tu at ICCV 2005

10. Metropolis-Hastings Algorithm Notes: Slides by Zhu, Dellaert and Tu at ICCV 2005

11. Metropolis-Hastings Algorithm Proposal Distribution Invariant Distribution Notes: Slides by Zhu, Dellaert and Tu at ICCV 2005

12. Reversible Jumps MCMC Many competing models to explain data –Need to explore this complicated state space Notes: Slides by Zhu, Dellaert and Tu at ICCV 2005

13. DDMCMC Motivation Notes: Slides by Zhu, Dellaert and Tu at ICCV 2005

14. DDMCMC Motivation Generative Model p(I|W)p(W) State Space

15. DDMCMC Motivation Generative Model p(I|W)p(W) State Space Discriminative Model q( w j | I ) Dramatically reduce search space by focusing sampling to highly probable states.

16. DDMCMC Framework Moves: –Node Creation –Node Deletion –Change Node Attributes

17. Transition Kernel Satisfies detailed balanced equation Full Transition Kernel

18. Convergence to p(W|I) Monotonically at a geometric rate

19. Criteria for Designing Transition Kernels

20. Image Generation Model Regions: Constant Intensity Textures Shading State of parse graph

21. 62 characters Faces 3 Regions

22. Uniform Designed to penalize high model complexity

23. Shape Prior Faces 3 Regions

24. Shape Prior: Text

25. Intensity Models

26. Intensity Model: Faces

27. Discriminative Cues Used Adaboost Trained –Face Detector –Text Detector Adaptive Binarization Cues Edge Cues –Canny at 3 scales Shape Affinity Cues Region Affinity Cues

28. Transition Kernel Design Remember

29. Possible Transitions 1.Birth/Death of a Face Node 2.Birth/Death of Text Node 3.Boundary Evolution 4.Split/Merge Region 5.Change node attributes

30. Face/Text Transitions

31. Region Transitions

32. Change Node Attributes

33. Basic Control Algorithm

35. Results

37. Comments Well motivated but very complicated approach to THE HOLY GRAIL problem in vision –Good global convergence results for inference with very minor dependence on initial W. –Extensible to larger set of primitives and pattern types. Many details of the algorithm are missing and it is hard to understand the motivation for choices of values for some parameters Unclear if the p(W|I)’s for configurations with different class compositions are comparable. Derek’s comment on Adaboost false positives and their failure to report their exact improvement No quantitative results/comparison to other algorithms and approaches –It should be possible to design a simple experiment to measure performance on recognition/detection/localization tasks.

38. Thank You