O BJ C UT M. Pawan Kumar Philip Torr Andrew Zisserman UNIVERSITY OF OXFORD.

Slides:



Advertisements
Similar presentations
Using Strong Shape Priors for Multiview Reconstruction Yunda SunPushmeet Kohli Mathieu BrayPhilip HS Torr Department of Computing Oxford Brookes University.
Advertisements

POSE–CUT Simultaneous Segmentation and 3D Pose Estimation of Humans using Dynamic Graph Cuts Mathieu Bray Pushmeet Kohli Philip H.S. Torr Department of.
Mean-Field Theory and Its Applications In Computer Vision1 1.
OBJ CUT & Pose Cut CVPR 05 ECCV 06
Combinatorial Optimization and Computer Vision Philip Torr.
Solving Markov Random Fields using Second Order Cone Programming Relaxations M. Pawan Kumar Philip Torr Andrew Zisserman.
Solving Markov Random Fields using Dynamic Graph Cuts & Second Order Cone Programming Relaxations M. Pawan Kumar, Pushmeet Kohli Philip Torr.
The Layout Consistent Random Field for detecting and segmenting occluded objects CVPR, June 2006 John Winn Jamie Shotton.
Linear Time Methods for Propagating Beliefs Min Convolution, Distance Transforms and Box Sums Daniel Huttenlocher Computer Science Department December,
Pose Estimation and Segmentation of People in 3D Movies Karteek Alahari, Guillaume Seguin, Josef Sivic, Ivan Laptev Inria, Ecole Normale Superieure ICCV.
Joint Optimisation for Object Class Segmentation and Dense Stereo Reconstruction Ľubor Ladický, Paul Sturgess, Christopher Russell, Sunando Sengupta, Yalin.
Agenda Introduction Bag-of-words models Visual words with spatial location Part-based models Discriminative methods Segmentation and recognition Recognition-based.
Part 4: Combined segmentation and recognition by Rob Fergus (MIT)
Learning to Combine Bottom-Up and Top-Down Segmentation Anat Levin and Yair Weiss School of CS&Eng, The Hebrew University of Jerusalem, Israel.
Carolina Galleguillos, Brian McFee, Serge Belongie, Gert Lanckriet Computer Science and Engineering Department Electrical and Computer Engineering Department.
ICCV 2007 tutorial Part III Message-passing algorithms for energy minimization Vladimir Kolmogorov University College London.
I Images as graphs Fully-connected graph – node for every pixel – link between every pair of pixels, p,q – similarity w ij for each link j w ij c Source:
Mixture of trees model: Face Detection, Pose Estimation and Landmark Localization Presenter: Zhang Li.
Human Pose detection Abhinav Golas S. Arun Nair. Overview Problem Previous solutions Solution, details.
An Analysis of Convex Relaxations (PART I) Minimizing Higher Order Energy Functions (PART 2) Philip Torr Work in collaboration with: Pushmeet Kohli, Srikumar.
LOCUS (Learning Object Classes with Unsupervised Segmentation) A variational approach to learning model- based segmentation. John Winn Microsoft Research.
GrabCut Interactive Image (and Stereo) Segmentation Carsten Rother Vladimir Kolmogorov Andrew Blake Antonio Criminisi Geoffrey Cross [based on Siggraph.
GrabCut Interactive Foreground Extraction using Iterated Graph Cuts Carsten Rother Vladimir Kolmogorov Andrew Blake Microsoft Research Cambridge-UK.
Graph-based image segmentation Václav Hlaváč Czech Technical University in Prague Faculty of Electrical Engineering Department of Cybernetics Prague, Czech.
GrabCut Interactive Image (and Stereo) Segmentation Joon Jae Lee Keimyung University Welcome. I will present Grabcut – an Interactive tool for foreground.
Simultaneous Segmentation and 3D Pose Estimation of Humans or Detection + Segmentation = Tracking? Philip H.S. Torr Pawan Kumar, Pushmeet Kohli, Matt Bray.
Robust Higher Order Potentials For Enforcing Label Consistency
Schedule Introduction Models: small cliques and special potentials Tea break Inference: Relaxation techniques:
ICCV Tutorial 2007 Philip Torr Papers, presentations and videos on web.....
An Analysis of Convex Relaxations M. Pawan Kumar Vladimir Kolmogorov Philip Torr for MAP Estimation.
P 3 & Beyond Solving Energies with Higher Order Cliques Pushmeet Kohli Pawan Kumar Philip H. S. Torr Oxford Brookes University CVPR 2007.
Improved Moves for Truncated Convex Models M. Pawan Kumar Philip Torr.
Efficiently Solving Convex Relaxations M. Pawan Kumar University of Oxford for MAP Estimation Philip Torr Oxford Brookes University.
The Layout Consistent Random Field for Recognizing and Segmenting Partially Occluded Objects By John Winn & Jamie Shotton CVPR 2006 presented by Tomasz.
Graph Cut based Inference with Co-occurrence Statistics Ľubor Ladický, Chris Russell, Pushmeet Kohli, Philip Torr.
What Energy Functions Can be Minimized Using Graph Cuts? Shai Bagon Advanced Topics in Computer Vision June 2010.
Simultaneous Segmentation and 3D Pose Estimation of Humans Philip H.S. Torr Pawan Kumar, Pushmeet Kohli, Matt Bray Oxford Brookes University Arasanathan.
Measuring Uncertainty in Graph Cut Solutions Pushmeet Kohli Philip H.S. Torr Department of Computing Oxford Brookes University.
Perceptual Organization: Segmentation and Optical Flow.
What, Where & How Many? Combining Object Detectors and CRFs
Graph-based Segmentation
Reconstructing Relief Surfaces George Vogiatzis, Philip Torr, Steven Seitz and Roberto Cipolla BMVC 2004.
MRFs and Segmentation with Graph Cuts Computer Vision CS 543 / ECE 549 University of Illinois Derek Hoiem 03/31/15.
Prakash Chockalingam Clemson University Non-Rigid Multi-Modal Object Tracking Using Gaussian Mixture Models Committee Members Dr Stan Birchfield (chair)
MRFs and Segmentation with Graph Cuts Computer Vision CS 543 / ECE 549 University of Illinois Derek Hoiem 02/24/10.
Object Stereo- Joint Stereo Matching and Object Segmentation Computer Vision and Pattern Recognition (CVPR), 2011 IEEE Conference on Michael Bleyer Vienna.
Discrete Optimization Lecture 2 – Part I M. Pawan Kumar Slides available online
Associative Hierarchical CRFs for Object Class Image Segmentation Ľubor Ladický 1 1 Oxford Brookes University 2 Microsoft Research Cambridge Based on the.
Efficient Discriminative Learning of Parts-based Models M. Pawan Kumar Andrew Zisserman Philip Torr
Associative Hierarchical CRFs for Object Class Image Segmentation
O BJ C UT M. Pawan Kumar Philip Torr Andrew Zisserman UNIVERSITY OF OXFORD.
Paper Reading Dalong Du Nov.27, Papers Leon Gu and Takeo Kanade. A Generative Shape Regularization Model for Robust Face Alignment. ECCV08. Yan.
Discussion of Pictorial Structures Pedro Felzenszwalb Daniel Huttenlocher Sicily Workshop September, 2006.
Category Independent Region Proposals Ian Endres and Derek Hoiem University of Illinois at Urbana-Champaign.
Inference for Learning Belief Propagation. So far... Exact methods for submodular energies Approximations for non-submodular energies Move-making ( N_Variables.
Object Recognition as Ranking Holistic Figure-Ground Hypotheses Fuxin Li and Joao Carreira and Cristian Sminchisescu 1.
Object Recognition by Integrating Multiple Image Segmentations Caroline Pantofaru, Cordelia Schmid, Martial Hebert ECCV 2008 E.
Efficient Belief Propagation for Image Restoration Qi Zhao Mar.22,2006.
Part 4: combined segmentation and recognition Li Fei-Fei.
Cutting Images: Graphs and Boundary Finding Computational Photography Derek Hoiem, University of Illinois 09/20/12 “The Double Secret”, Magritte.
Image segmentation.
MRFs and Segmentation with Graph Cuts Computer Vision CS 543 / ECE 549 University of Illinois Derek Hoiem 03/27/12.
GrabCut Interactive Foreground Extraction using Iterated Graph Cuts Carsten Rother Vladimir Kolmogorov Andrew Blake Microsoft Research Cambridge-UK.
LOCUS: Learning Object Classes with Unsupervised Segmentation
Nonparametric Semantic Segmentation
Learning to Combine Bottom-Up and Top-Down Segmentation
“The Truth About Cats And Dogs”
Learning Layered Motion Segmentations of Video
PRAKASH CHOCKALINGAM, NALIN PRADEEP, AND STAN BIRCHFIELD
“Traditional” image segmentation
Presentation transcript:

O BJ C UT M. Pawan Kumar Philip Torr Andrew Zisserman UNIVERSITY OF OXFORD

Aim Given an image, to segment the object Segmentation should (ideally) be shaped like the object e.g. cow-like obtained efficiently in an unsupervised manner able to handle self-occlusion Segmentation Object Category Model Cow Image Segmented Cow

Challenges Self Occlusion Intra-Class Shape Variability Intra-Class Appearance Variability

Motivation Magic Wand Current methods require user intervention Object and background seed pixels (Boykov and Jolly, ICCV 01) Bounding Box of object (Rother et al. SIGGRAPH 04) Cow Image Object Seed Pixels

Motivation Magic Wand Current methods require user intervention Object and background seed pixels (Boykov and Jolly, ICCV 01) Bounding Box of object (Rother et al. SIGGRAPH 04) Cow Image Object Seed Pixels Background Seed Pixels

Motivation Magic Wand Current methods require user intervention Object and background seed pixels (Boykov and Jolly, ICCV 01) Bounding Box of object (Rother et al. SIGGRAPH 04) Segmented Image

Motivation Magic Wand Current methods require user intervention Object and background seed pixels (Boykov and Jolly, ICCV 01) Bounding Box of object (Rother et al. SIGGRAPH 04) Cow Image Object Seed Pixels Background Seed Pixels

Motivation Magic Wand Current methods require user intervention Object and background seed pixels (Boykov and Jolly, ICCV 01) Bounding Box of object (Rother et al. SIGGRAPH 04) Segmented Image

Problem Manually intensive Segmentation is not guaranteed to be object-like Non Object-like Segmentation Motivation

Our Method Combine object detection with segmentation –Borenstein and Ullman, ECCV 02 –Leibe and Schiele, BMVC 03 Incorporate global shape priors in MRF Detection provides – Object Localization – Global shape priors Automatically segments the object –Note our method is completely generic –Applicable to any object category model

Outline Problem Formulation Form of Shape Prior Optimization Results

Problem Labelling m over the set of pixels D Shape prior provided by parameter Energy E (m, ) = x (D|m x )+ x (m x | ) + xy (m x,m y )+ (D|m x,m y ) Unary terms –Likelihood based on colour –Unary potential based on distance from Pairwise terms –Prior –Contrast term Find best labelling m* = arg min w i E (m, i ) –w i is the weight for sample i Unary termsPairwise terms

MRF Probability for a labelling consists of Likelihood Unary potential based on colour of pixel Prior which favours same labels for neighbours (pairwise potentials) D (pixels) m (labels) Image Plane x y mxmx mymy Unary Potential x (D|m x ) Pairwise Potential xy (m x, m y )

Example Cow Image Object Seed Pixels Background Seed Pixels Prior x … y … … … x … y … … … x (D|obj) x (D|bkg) xy (m x,m y ) Likelihood Ratio (Colour)

Example Cow Image Object Seed Pixels Background Seed Pixels Prior Likelihood Ratio (Colour)

Contrast-Dependent MRF Probability of labelling in addition has Contrast term which favours boundaries to lie on image edges D (pixels) m (labels) Image Plane Contrast Term (D|m x,m y ) x y mxmx mymy

Example Cow Image Object Seed Pixels Background Seed Pixels Prior + Contrast x … y … … … x … y … … … Likelihood Ratio (Colour) x (D|obj) x (D|bkg) xy (m x,m y )+ xy (D|m x,m y )

Example Cow Image Object Seed Pixels Background Seed Pixels Prior + Contrast Likelihood Ratio (Colour)

Our Model Probability of labelling in addition has Unary potential which depend on distance from (shape parameter) D (pixels) m (labels) (shape parameter) Image Plane Object Category Specific MRF x y mxmx mymy Unary Potential x (m x | )

Example Cow Image Object Seed Pixels Background Seed Pixels Prior + Contrast Distance from Shape Prior

Example Cow Image Object Seed Pixels Background Seed Pixels Prior + Contrast Likelihood + Distance from Shape Prior

Example Cow Image Object Seed Pixels Background Seed Pixels Prior + Contrast Likelihood + Distance from Shape Prior

Outline Problem Formulation –Energy E (m, ) = x (D|m x )+ x (m x | ) + xy (m x,m y )+ (D|m x,m y ) Form of Shape Prior Optimization Results

Layered Pictorial Structures (LPS) Generative model Composition of parts + spatial layout Layer 2 Layer 1 Parts in Layer 2 can occlude parts in Layer 1 Spatial Layout (Pairwise Configuration)

Layer 2 Layer 1 Transformations 1 P( 1 ) = 0.9 Cow Instance Layered Pictorial Structures (LPS)

Layer 2 Layer 1 Transformations 2 P( 2 ) = 0.8 Cow Instance Layered Pictorial Structures (LPS)

Layer 2 Layer 1 Transformations 3 P( 3 ) = 0.01 Unlikely Instance Layered Pictorial Structures (LPS)

LPS for Detection Learning – Learnt automatically using a set of videos – Part correspondence using Shape Context Shape Context Matching Multiple Shape Exemplars

LPS for Detection Detection – Putative parts found using tree cascade of classifiers (x,y)

LPS for Detection MRF over parts Labels represent putative poses Prior (pairwise potential) - Robust Truncated Model Match LPS by obtaining MAP configuration Potts Model Linear ModelQuadratic Model

LPS for Detection Efficient Belief Propagation j i k Likelihood i (x i ) tree cascade of classifiers Prior ij (x i,x j ) f ij (x i,x j ), if x i C i (x j ) ij, otherwise Pr(x) i (x i ) ij (x i,x j ) xixi xjxj xkxk ij jk ki i j k Messages m j->i

LPS for Detection Efficient Belief Propagation j i k Likelihood i (x i ) tree cascade of classifiers Prior ij (x i,x j ) f ij (x i,x j ), if x i C i (x j ) ij, otherwise Pr(x) i (x i ) ij (x i,x j ) xixi xjxj xkxk Messages calculated as

LPS for Detection Efficient Generalized Belief Propagation j i k Likelihood i (x i ) tree cascade of classifiers Prior ij (x i,x j ) f ij (x i,x j ), if x i C i (x j ) ij, otherwise Pr(x) i (x i ) ij (x i,x j ) xixi xjxj xkxk ij jk ki i j k Messages m k->ij ijk

LPS for Detection Efficient Generalized Belief Propagation j i k Likelihood i (x i ) tree cascade of classifiers Prior ij (x i,x j ) f ij (x i,x j ), if x i C i (x j ) ij, otherwise Pr(x) i (x i ) ij (x i,x j ) xixi xjxj xkxk Messages calculated as

LPS for Detection Second Order Cone Programming Relaxations j i k Likelihood i (x i ) tree cascade of classifiers Prior ij (x i,x j ) f ij (x i,x j ), if x i C i (x j ) ij, otherwise Pr(x) i (x i ) ij (x i,x j ) xixi xjxj xkxk

LPS for Detection Second Order Cone Programming Relaxations j i k Likelihood i (x i ) tree cascade of classifiers Prior ij (x i,x j ) f ij (x i,x j ), if x i C i (x j ) ij, otherwise Pr(x) i (x i ) ij (x i,x j ) m - Concatenation of all binary vectors l - Likelihood vector P - Prior matrix

LPS for Detection Second Order Cone Programming Relaxations j i k

LPS for Detection Second Order Cone Programming Relaxations j i k

LPS for Detection Second Order Cone Programming Relaxations j i k

Outline Problem Formulation Form of Shape Prior Optimization Results

Optimization Given image D, find best labelling as m* = arg max p(m|D) Treat LPS parameter as a latent (hidden) variable EM framework –E : sample the distribution over –M : obtain the labelling m

E-Step Given initial labelling m, determine p( | m,D) Problem Efficiently sampling from p( | m,D) Solution We develop efficient sum-product Loopy Belief Propagation (LBP) for matching LPS. Similar to efficient max-product LBP for MAP estimate

Results Different samples localize different parts well. We cannot use only the MAP estimate of the LPS.

M-Step Given samples from p( |m,D), get new labelling m new Sample i provides –Object localization to learn RGB distributions of object and background –Shape prior for segmentation Problem –Maximize expected log likelihood using all samples –To efficiently obtain the new labelling

M-Step Cow Image Shape 1 w 1 = P( 1 |m,D) RGB Histogram for Object RGB Histogram for Background

Cow Image M-Step 1 Image Plane D (pixels) m (labels) Best labelling found efficiently using a Single Graph Cut Shape 1 w 1 = P( 1 |m,D)

Segmentation using Graph Cuts x … y ……… z …… Obj Bkg Cut x (D|bkg) + x (bkg| ) m z (D|obj) + z (obj| ) xy (m x,m y )+ xy (D|m x,m y )

Segmentation using Graph Cuts x … y ……… z …… Obj Bkg m

M-Step Cow Image RGB Histogram for Background RGB Histogram for Object Shape 2 w 2 = P( 2 |m,D)

M-Step Cow Image 2 Image Plane D (pixels) m (labels) Best labelling found efficiently using a Single Graph Cut Shape 2 w 2 = P( 2 |m,D)

M-Step 2 Image Plane 1 w1w1 + w 2 + …. Best labelling found efficiently using a Single Graph Cut m* = arg min w i E (m, i )

Outline Problem Formulation Form of Shape Prior Optimization Results

SegmentationImage Results Using LPS Model for Cow

In the absence of a clear boundary between object and background SegmentationImage Results Using LPS Model for Cow

SegmentationImage Results Using LPS Model for Cow

SegmentationImage Results Using LPS Model for Cow

SegmentationImage Results Using LPS Model for Horse

SegmentationImage Results Using LPS Model for Horse

Our MethodLeibe and SchieleImage Results

Appearance ShapeShape+Appearance Results Without x (D|m x ) Without x (m x | )

Conclusions –New model for introducing global shape prior in MRF –Method of combining detection and segmentation –Efficient LBP for detecting articulated objects Future Work –Other shape parameters need to be explored –Method needs to be extended to handle multiple visual aspects