scene labeling using beam search under mutex constraints anirban roy and sinisa todorovic
DESCRIPTION
Scene Labeling Using Beam Search Under Mutex Constraints Anirban Roy and Sinisa Todorovic. Beam Search for Solving QP Results Acknowledgment. Problem and Motivation Approach Extracting superpixels Incorporating mutex in the standard CRF formulation Formulating CRF inference as QP - PowerPoint PPT PresentationTRANSCRIPT
IEEE 2014 Conference on Computer Vision and Pattern
Recognition
Beam Search for Solving QPBeam Search for Solving QP
Results Results
Acknowledgment Acknowledgment
Problem and MotivationProblem and Motivation
ApproachApproach1.1. Extracting superpixels Extracting superpixels 2.2. Incorporating mutex in the standard CRF formulationIncorporating mutex in the standard CRF formulation3.3. Formulating CRF inference as QPFormulating CRF inference as QP4.4. Beam search for solving QPBeam search for solving QP5.5. Learning – piecewiseLearning – piecewise
How to Specify CRF Energy?How to Specify CRF Energy?
CRF Inference as QPCRF Inference as QP
Specifying Mutex ConstraintsSpecifying Mutex Constraints
Scene Labeling Using Beam Search Under Mutex Constraints Anirban Roy and Sinisa Todorovic
Input Image Semantic segmentation with Mutex
Semantic segmentation without Mutex
Mutex violations
Appearance features of the superpixels
Smoothnessand Context
∞
Pixelwise accuracy(%)
can be arbitrary
Assignment vector
Matrix of CRF potentials
NSF RI 1302700
MUTual EXclusion = (object, object, relationship)
State: label assignmentHeuristic function:
Score:Superpixel Class label
Method MSRC Test timeGalleguillos et al. CVPR 10 70.4 N/A
Gould et al. ICCV 09 76.4 N/APayet et al. PAMI 12 82.9 30-32s
Krahenbuhl et al. NIPS 12 86.0 0.2sYao et al. CVPR 12 86.5 N/A
Zhang et al. CVPR 12 87.0 N/AOurs 91.5 0.8s
Method StanfordBackground
Gould et al. ICCV 09 76.4Munoz et al. ECCV 10 76.9Singh et al. CVPR 13 74.1
Ours 81.1
Matrix of mutex constraints
maximum score
next state previous state
must be