1 / 213

MAP Estimation Algorithms in

MAP Estimation Algorithms in. Computer Vision - Part I. M. Pawan Kumar, University of Oxford Pushmeet Kohli, Microsoft Research. Aim of the Tutorial. Description of some successful algorithms Computational issues Enough details to implement Some proofs will be skipped :-(

may-bartlett
Download Presentation

MAP Estimation Algorithms in

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. MAP Estimation Algorithms in Computer Vision - Part I M. Pawan Kumar, University of Oxford Pushmeet Kohli, Microsoft Research

  2. Aim of the Tutorial • Description of some successful algorithms • Computational issues • Enough details to implement • Some proofs will be skipped :-( • But references to them will be given :-)

  3. A Vision Application Binary Image Segmentation How ? Cost function Models our knowledge about natural images Optimize cost function to obtain the segmentation

  4. A Vision Application Binary Image Segmentation Graph G = (V,E) Object - white, Background - green/grey Each vertex corresponds to a pixel Edges define a 4-neighbourhood grid graph Assign a label to each vertex from L = {obj,bkg}

  5. A Vision Application Binary Image Segmentation Graph G = (V,E) Object - white, Background - green/grey Per Vertex Cost Cost of a labelling f : V  L Cost of label ‘bkg’ high Cost of label ‘obj’ low

  6. A Vision Application Binary Image Segmentation Graph G = (V,E) Object - white, Background - green/grey Per Vertex Cost Cost of a labelling f : V  L Cost of label ‘bkg’ low Cost of label ‘obj’ high UNARY COST

  7. A Vision Application Binary Image Segmentation Graph G = (V,E) Object - white, Background - green/grey Per Edge Cost Cost of a labelling f : V  L Cost of same label low Cost of different labels high

  8. A Vision Application Binary Image Segmentation Graph G = (V,E) Object - white, Background - green/grey Per Edge Cost Cost of a labelling f : V  L Cost of same label high PAIRWISE COST Cost of different labels low

  9. A Vision Application Binary Image Segmentation Graph G = (V,E) Object - white, Background - green/grey Problem: Find the labelling with minimum cost f*

  10. A Vision Application Binary Image Segmentation Graph G = (V,E) Problem: Find the labelling with minimum cost f*

  11. Another Vision Application Object Detection using Parts-based Models How ? Once again, by defining a good cost function

  12. Another Vision Application Object Detection using Parts-based Models H T 1 L1 L2 L3 L4 Graph G = (V,E) Each vertex corresponds to a part - ‘Head’, ‘Torso’, ‘Legs’ Edges define a TREE Assign a label to each vertex from L = {positions}

  13. Another Vision Application Object Detection using Parts-based Models H T 2 L1 L2 L3 L4 Graph G = (V,E) Each vertex corresponds to a part - ‘Head’, ‘Torso’, ‘Legs’ Edges define a TREE Assign a label to each vertex from L = {positions}

  14. Another Vision Application Object Detection using Parts-based Models H T 3 L1 L2 L3 L4 Graph G = (V,E) Each vertex corresponds to a part - ‘Head’, ‘Torso’, ‘Legs’ Edges define a TREE Assign a label to each vertex from L = {positions}

  15. Another Vision Application Object Detection using Parts-based Models H T 3 L1 L2 L3 L4 Graph G = (V,E) Cost of a labelling f : V  L Unary cost : How well does part match image patch? Pairwise cost : Encourages valid configurations Find best labelling f*

  16. Another Vision Application Object Detection using Parts-based Models H T 3 L1 L2 L3 L4 Graph G = (V,E) Cost of a labelling f : V  L Unary cost : How well does part match image patch? Pairwise cost : Encourages valid configurations Find best labelling f*

  17. Yet Another Vision Application Stereo Correspondence Disparity Map How ? Minimizing a cost function

  18. Yet Another Vision Application Stereo Correspondence Graph G = (V,E) Vertex corresponds to a pixel Edges define grid graph L = {disparities}

  19. Yet Another Vision Application Stereo Correspondence Cost of labelling f : Unary cost + Pairwise Cost Find minimum cost f*

  20. The General Problem 1 2 b c Graph G = ( V, E ) 1 Discrete label set L = {1,2,…,h} a d 3 Assign a label to each vertex f: V  L f e 2 2 Cost of a labelling Q(f) Unary Cost Pairwise Cost Find f* = arg min Q(f)

  21. Outline • Problem Formulation • Energy Function • MAP Estimation • Computing min-marginals • Reparameterization • Belief Propagation • Tree-reweighted Message Passing

  22. Energy Function Label l1 Label l0 Vb Vc Vd Va Db Dc Dd Da Random Variables V= {Va, Vb, ….} Labels L= {l0, l1, ….} Data D Labelling f: {a, b, …. }  {0,1, …}

  23. Energy Function 6 3 2 4 Label l1 Label l0 5 3 7 2 Vb Vc Vd Va Db Dc Dd Da Easy to minimize Q(f) = ∑a a;f(a) Neighbourhood Unary Potential

  24. Energy Function 6 3 2 4 Label l1 Label l0 5 3 7 2 Vb Vc Vd Va Db Dc Dd Da E : (a,b)  E iff Va and Vb are neighbours E = { (a,b) , (b,c) , (c,d) }

  25. Energy Function 0 6 1 3 2 0 4 Label l1 1 2 3 4 1 1 Label l0 1 0 5 0 3 7 2 Vb Vc Vd Va Db Dc Dd Da Pairwise Potential Q(f) = ∑a a;f(a) +∑(a,b) ab;f(a)f(b)

  26. Energy Function 0 6 1 3 2 0 4 Label l1 1 2 3 4 1 1 Label l0 1 0 5 0 3 7 2 Vb Vc Vd Va Db Dc Dd Da Q(f; ) = ∑a a;f(a) +∑(a,b) ab;f(a)f(b) Parameter

  27. Outline • Problem Formulation • Energy Function • MAP Estimation • Computing min-marginals • Reparameterization • Belief Propagation • Tree-reweighted Message Passing

  28. MAP Estimation 0 6 1 3 2 0 4 Label l1 1 2 3 4 1 1 Label l0 1 0 5 0 3 7 2 Vb Vc Vd Va Q(f; ) = ∑a a;f(a) + ∑(a,b) ab;f(a)f(b)

  29. MAP Estimation 0 6 1 3 2 0 4 Label l1 1 2 3 4 1 1 Label l0 1 0 5 0 3 7 2 Vb Vc Vd Va Q(f; ) = ∑a a;f(a) + ∑(a,b) ab;f(a)f(b) 2 + 1 + 2 + 1 + 3 + 1 + 3 = 13

  30. MAP Estimation 0 6 1 3 2 0 4 Label l1 1 2 3 4 1 1 Label l0 1 0 5 0 3 7 2 Vb Vc Vd Va Q(f; ) = ∑a a;f(a) + ∑(a,b) ab;f(a)f(b)

  31. MAP Estimation 0 6 1 3 2 0 4 Label l1 1 2 3 4 1 1 Label l0 1 0 5 0 3 7 2 Vb Vc Vd Va Q(f; ) = ∑a a;f(a) + ∑(a,b) ab;f(a)f(b) 5 + 1 + 4 + 0 + 6 + 4 + 7 = 27

  32. MAP Estimation 0 6 1 3 2 0 4 Label l1 1 2 3 4 1 1 Label l0 1 0 5 0 3 7 2 Vb Vc Vd Va q* = min Q(f; ) = Q(f*; ) Q(f; ) = ∑a a;f(a) + ∑(a,b) ab;f(a)f(b) f* = arg min Q(f; )

  33. MAP Estimation f* = {1, 0, 0, 1} 16 possible labellings q* = 13

  34. Computational Complexity Segmentation 2|V| |V| = number of pixels ≈ 320 * 480 = 153600

  35. Computational Complexity Detection |L||V| |L| = number of pixels ≈ 153600

  36. Computational Complexity Stereo |L||V| |V| = number of pixels ≈ 153600 Can we do better than brute-force? MAP Estimation is NP-hard !!

  37. Computational Complexity Stereo |L||V| |V| = number of pixels ≈ 153600 Exact algorithms do exist for special cases Good approximate algorithms for general case But first … two important definitions

  38. Outline • Problem Formulation • Energy Function • MAP Estimation • Computing min-marginals • Reparameterization • Belief Propagation • Tree-reweighted Message Passing

  39. Min-Marginals 0 6 1 3 2 0 4 Label l1 1 2 3 4 1 1 Label l0 1 0 5 0 3 7 2 Vb Vc Vd Va Not a marginal (no summation) such that f(a) = i f* = arg min Q(f; ) Min-marginal qa;i

  40. Min-Marginals qa;0 = 15 16 possible labellings

  41. Min-Marginals qa;1 = 13 16 possible labellings

  42. Min-Marginals and MAP • Minimum min-marginal of any variable = • energy of MAP labelling qa;i mini ) mini ( such that f(a) = i minf Q(f; ) Va has to take one label minf Q(f; )

  43. Summary Energy Function Q(f; ) = ∑a a;f(a) + ∑(a,b) ab;f(a)f(b) MAP Estimation f* = arg min Q(f; ) Min-marginals s.t. f(a) = i qa;i = min Q(f; )

  44. Outline • Problem Formulation • Reparameterization • Belief Propagation • Tree-reweighted Message Passing

  45. Reparameterization 2 + - 2 2 0 4 1 1 2 + - 2 5 0 2 Vb Va Add a constant to all a;i Subtract that constant from all b;k

  46. Reparameterization 2 + - 2 2 0 4 1 1 2 + - 2 5 0 2 Vb Va Add a constant to all a;i Subtract that constant from all b;k Q(f; ’) = Q(f; )

  47. Reparameterization + 3 - 3 2 0 4 - 3 1 1 5 0 2 Vb Va Add a constant to one b;k Subtract that constant from ab;ik for all ‘i’

  48. Reparameterization + 3 - 3 2 0 4 - 3 1 1 5 0 2 Vb Va Add a constant to one b;k Subtract that constant from ab;ik for all ‘i’ Q(f; ’) = Q(f; )

  49. 3 1 3 1 3 0 1 1 1 2 2 4 2 4 2 4 0 1 2 1 5 0 5 5 2 2 2 Vb Vb Vb Va Va Va Reparameterization - 4 + 4 + 1 - 4 - 2 - 1 + 1 - 2 - 4 + 1 - 2 + 2 + Mab;k + Mba;i ’b;k = b;k ’a;i = a;i Q(f; ’) = Q(f; ) ’ab;ik = ab;ik - Mab;k - Mba;i

  50. Equivalently 2 + - 2 2 0 4 + Mba;i ’a;i = a;i 1 1 + Mab;k 2 + - 2 5 0 2 Vb Va ’ab;ik = ab;ik - Mab;k - Mba;i Reparameterization ’ is a reparameterization of , iff ’   Q(f; ’) = Q(f; ), for all f Kolmogorov, PAMI, 2006 ’b;k = b;k

More Related