en
0.25
0.5
0.75
1.25
1.5
1.75
2
Generic Cuts: An Efficient Algorithm for Optimal Inference in Higher Order MRF-MAP
Published on Nov 12, 20124240 Views
We propose a new algorithm called Generic Cuts for computing optimal solutions to 2 label MRF-MAP problems with higher order clique potentials satisfying submodularity. The algorithm runs in time O(2
Related categories
Chapter list
GENERIC CUTS: AN EFFICIENT ALGORITHM FOR OPTIMAL INFERENCE IN HIGHER ORDER MRF-MAP00:00
MRF-MAP INFERENCE (1)00:13
MRF-MAP INFERENCE (2)00:43
HIGHER ORDER MRF – WHY?01:01
POSSIBLE INFERENCE METHODS01:53
LIMITATIONS02:24
MAIN CONTRIBUTION03:57
GADGET04:57
EDGE CAPACITY – DUAL FEASIBILITY CONSTRAINT (1)06:06
EDGE CAPACITY – DUAL FEASIBILITY CONSTRAINT (2)07:00
HOW DOES SUBMODULARITY EFFECTS US - MAXFLOW?08:15
HOW DOES SUBMODULARITY EFFECTS US - MINCUT?08:50
COMBINATORIAL PROPERTIES OF THE FRAMEWORK10:15
COMPARISON11:43
COMPARISON -ENERGY , CLIQUE SIZE=412:49
COMPARISON -TIME (MS) , CLIQUE SIZE=413:19
COMPARISON – CLIQUE SIZE V/S TIME (MS)13:27
GENERALIZATIONS13:58
SUMMARY14:12