Download presentation

Presentation is loading. Please wait.

Published byJordan Rose Modified over 4 years ago

1
Algorithms for MAP estimation in Markov Random Fields Vladimir Kolmogorov University College London Tutorial at GDR (Optimisation Discrète, Graph Cuts et Analyse d'Images) Paris, 29 November 2005 Note: these slides contain animation

2
Energy function p q unary terms (data) pairwise terms (coherence) - x p are discrete variables (for example, x p {0,1}) - p ( ) are unary potentials - pq (, ) are pairwise potentials

3
Minimisation algorithms Min Cut / Max Flow [Ford&Fulkerson 56] [Grieg, Porteous, Seheult 89] : non-iterative (binary variables) [Boykov, Veksler, Zabih 99] : iterative - alpha-expansion, alpha-beta swap, … (multi-valued variables) + If applicable, gives very accurate results – Can be applied to a restricted class of functions BP – Max-product Belief Propagation [Pearl 86] + Can be applied to any energy function – In vision results are usually worse than that of graph cuts – Does not always converge TRW - Max-product Tree-reweighted Message Passing [Wainwright, Jaakkola, Willsky 02], [Kolmogorov 05] + Can be applied to any energy function + For stereo finds lower energy than graph cuts + Convergence guarantees for the algorithm in [Kolmogorov 05]

4
Main idea: LP relaxation Goal: Minimize energy E(x) under constraints x p {0,1} In general, NP-hard problem! Relax discreteness constraints: allow x p [0,1] Results in linear program. Can be solved in polynomial time! Energy function with discrete variables LP relaxation E E E tight not tight

5
Solving LP relaxation Too large for general purpose LP solvers (e.g. interior point methods) Solve dual problem instead of primal: –Formulate lower bound on the energy –Maximize this bound –When done, solves primal problem (LP relaxation) Two different ways to formulate lower bound –Via posiforms: leads to maxflow algorithm –Via convex combination of trees: leads to tree-reweighted message passing Lower bound on the energy function E Energy function with discrete variables E E LP relaxation

6
Notation and Preliminaries

7
Energy function - visualisation 0 4 0 1 3 02 5 node p edge (p,q) node q label 0 label 1 0

8
0 4 0 1 3 02 5 node p edge (p,q) node q label 0 label 1 Energy function - visualisation 0 vector of all parameters

9
0 0 4 4 1 1 2 5 0 0 + 1 Reparameterisation

10
0 0 3 4 1 0 2 5 Definition. is a reparameterisation of if they define the same energy: 4 -1 1 -1 0 +1 Maxflow, BP and TRW perform reparameterisations 1

11
Part I: Lower bound via posiforms ( maxflow algorithm)

12
non-negative - lower bound on the energy: maximize Lower bound via posiforms [Hammer, Hansen, Simeone84]

13
Maximisation algorithm? –Consider functions of binary variables only Maximising lower bound for submodular functions –Definition of submodular functions –Overview of min cut/max flow –Reduction to max flow –Global minimum of the energy Maximising lower bound for non-submodular functions –Reduction to max flow More complicated graph –Part of optimal solution Outline of part I

14
Definition: E is submodular if every pairwise term satisfies Can be converted to canonical form: Submodular functions of binary variables 2 1234 1 0 00 5 zero cost

15
Overview of min cut/max flow

16
Min Cut problem source sink 2 1 1 2 3 4 5 Directed weighted graph

17
Min Cut problem sink 2 1 1 2 3 4 5 S = {source, node 1} T = {sink, node 2, node 3} Cut: source

18
Min Cut problem sink 2 1 1 2 3 4 5 S = {source, node 1} T = {sink, node 2, node 3} Cut: Task: Compute cut with minimum cost Cost(S,T) = 1 + 1 = 2 source

19
sink 2 1 1 2 3 4 5 source Maxflow algorithm value(flow)=0

20
Maxflow algorithm sink 2 1 1 2 3 4 5 value(flow)=0 source

21
Maxflow algorithm sink 1 1 0 3 3 4 4 value(flow)=1 source

22
Maxflow algorithm sink 1 1 0 3 3 4 4 value(flow)=1 source

23
Maxflow algorithm sink 1 0 0 3 4 3 3 value(flow)=2 source

24
Maxflow algorithm sink 1 0 0 3 4 3 3 value(flow)=2 source

25
value(flow)=2 sink 1 0 0 3 4 3 3 source Maxflow algorithm

26
Maximising lower bound for submodular functions: Reduction to maxflow

27
2 1234 1 0 00 5 sink 2 1 1 2 3 4 5 source value(flow)=0 0 Maxflow algorithm and reparameterisation

28
sink 2 1 1 2 3 4 5 value(flow)=0 2 1234 1 0 00 5 0 source Maxflow algorithm and reparameterisation

29
sink 1 1 0 3 3 4 4 value(flow)=1 1 0334 1 0 00 4 1 source Maxflow algorithm and reparameterisation

30
sink 1 1 0 3 3 4 4 value(flow)=1 1 0334 1 0 00 4 1 source Maxflow algorithm and reparameterisation

31
sink 1 0 0 3 4 3 3 value(flow)=2 1 0343 0 0 00 3 2 source Maxflow algorithm and reparameterisation

32
sink 1 0 0 3 4 3 3 value(flow)=2 1 0343 0 0 00 3 2 source Maxflow algorithm and reparameterisation

33
value(flow)=2 0 0 0 0 minimum of the energy: 2 0 sink 1 0 0 3 4 3 3 source Maxflow algorithm and reparameterisation

34
Maximising lower bound for non-submodular functions

35
Arbitrary functions of binary variables Can be solved via maxflow [Boros,Hammer,Sun91] –Specially constructed graph Gives solution to LP relaxation: for each node x p {0, 1/2, 1} E LP relaxation non-negative maximize

36
Arbitrary functions of binary variables 0 1 0 1 11/2 Part of optimal solution [Hammer, Hansen, Simeone84]

37
Part II: Lower bound via convex combination of trees ( tree-reweighted message passing)

38
Goal: compute minimum of the energy for In general, intractable! Obtaining lower bound: –Split into several components: –Compute minimum for each component: –Combine to get a bound on Use trees! Convex combination of trees [Wainwright, Jaakkola, Willsky 02]

39
graph tree T lower bound on the energy maximize Convex combination of trees (contd)

40
TRW algorithms Goal: find reparameterisation maximizing lower bound Apply sequence of different reparameterisation operations: –Node averaging –Ordinary BP on trees Order of operations? –Affects performance dramatically Algorithms: –[Wainwright et al. 02]: parallel schedule May not converge –[Kolmogorov05]: specific sequential schedule Lower bound does not decrease, convergence guarantees

41
Node averaging 0 1 4 0

42
2 0.5 2

43
Send messages –Equivalent to reparameterising node and edge parameters Two passes (forward and backward) Belief propagation (BP) on trees

44
3 0 Key property (Wainwright et al.): Upon termination p gives min-marginals for node p:

45
TRW algorithm of Wainwright et al. with tree-based updates (TRW-T) Run BP on all trees Average all nodes If converges, gives (local) maximum of lower bound Not guaranteed to converge. Lower bound may go down.

46
Sequential TRW algorithm (TRW-S) [Kolmogorov05] Run BP on all trees containing p Average node p Pick node p

47
Main property of TRW-S Theorem: lower bound never decreases. Proof sketch: 0 1 4 0

48
Main property of TRW-S 2 0.5 2 Theorem: lower bound never decreases. Proof sketch:

49
TRW-S algorithm Particular order of averaging and BP operations Lower bound guaranteed not to decrease There exists limit point that satisfies weak tree agreement condition Efficiency?

50
Average node p Pick node p inefficient? Efficient implementation Run BP on all trees containing p

51
Efficient implementation Key observation: Node averaging operation preserves messages oriented towards this node Reuse previously passed messages! Need a special choice of trees: –Pick an ordering of nodes –Trees: monotonic chains 456 789 123

52
Efficient implementation 456 789 123 Algorithm: –Forward pass: process nodes in the increasing order pass messages from lower neighbours –Backward pass: do the same in reverse order Linear running time of one iteration

53
Efficient implementation 456 789 123 Algorithm: –Forward pass: process nodes in the increasing order pass messages from lower neighbours –Backward pass: do the same in reverse order Linear running time of one iteration

54
Memory requirements Additional advantage of TRW-S: –Needs only half as much memory as standard message passing! –Similar observation for bipartite graphs and parallel schedule was made in [Felzenszwalb&Huttenlocher04] standard message passing TRW-S

55
Experimental results: binary segmentation (GrabCut) Time Energy average over 50 instances

56
Experimental results: stereo left image ground truth BP TRW-S

57
Experimental results: stereo

58
Summary MAP estimation algorithms are based on LP relaxation –Maximize lower bound Two ways to formulate lower bound Via posiforms: leads to maxflow algorithm –Polynomial time solution –But: applicable for restricted energies (e.g. binary variables) Submodular functions: global minimum Non-submodular functions: part of optimal solution Via convex combination of trees: leads to TRW algorithm –Convergence in the limit (for TRW-S) –Applicable to arbitrary energy function Graph cuts vs. TRW: –Accuracy: similar –Generality: TRW is more general –Speed: for stereo TRW is currently 2-5 times slower. But: 3 vs. 50 years of research! More suitable for parallel implementation (GPU? Hardware?)

59
Discrete vs. continuous functionals Continuous formulation (Geodesic active contours) Maxflow algorithm –Global minimum, polynomial-time Metrication artefacts? Level sets –Numerical stability? Geometrically motivated –Invariant under rotation Discrete formulation (Graph cuts)

60
Geo-cuts Continuous functional Construct graph such that for smooth contours C Class of continuous functionals? [Boykov&Kolmogorov03], [Kolmogorov&Boykov05]: –Geometric length/area (e.g. Riemannian) –Flux of a given vector field –Regional term

Similar presentations

OK

Discrete Optimization Lecture 3 – Part 1 M. Pawan Kumar Slides available online

Discrete Optimization Lecture 3 – Part 1 M. Pawan Kumar Slides available online

© 2018 SlidePlayer.com Inc.

All rights reserved.

To make this website work, we log user data and share it with processors. To use this website, you must agree to our Privacy Policy, including cookie policy.

Ads by Google

Ppt on non biodegradable wastewater Ppt on ideal gas law problems Ppt on tourism in karnataka Ppt on reverse mortgage Ppt on air powered car Pps to ppt online Ppt on buildings paintings and books Ppt on grammar translation method Ppt on chapter 3 atoms and molecules animation Ppt on nuclear power plant in india