DocumentCode :
2917573
Title :
Exhaustive family of energies minimizable exactly by a graph cut
Author :
Charpiat, Guillaume
Author_Institution :
INRIA, Sophia-Antipolis, France
fYear :
2011
fDate :
20-25 June 2011
Firstpage :
1849
Lastpage :
1856
Abstract :
Graph cuts are widely used in many fields of computer vision in order to minimize in small polynomial time complexity certain classes of energies. These specific classes depend on the way chosen to build the graphs representing the problems to solve. We study here all possible ways of building graphs and the associated energies minimized, leading to the exhaustive family of energies minimizable exactly by a graph cut. To do this, we consider the issue of coding pixel labels as states of the graph, i.e. the choice of state interpretations. The family obtained comprises many new classes, in particular energies that do not satisfy the submodularity condition, including energies that are even not permuted-submodular. A generating subfamily is studied in details, in particular we propose a canonical form to represent Markov random fields, which proves useful to recognize energies in this subfamily in linear complexity almost surely, and then to build the associated graph in quasilinear time. A few experiments are performed, to illustrate the new possibilities offered.
Keywords :
Markov processes; computational complexity; computer vision; graph theory; image coding; random processes; Markov random field; canonical form; computer vision; energies minimization; energy recognition; graph cut; linear complexity; pixel label coding; polynomial time complexity; quasilinear time; Equations; Frequency modulation; Labeling; Markov processes; Minimization; Silicon; Writing;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Computer Vision and Pattern Recognition (CVPR), 2011 IEEE Conference on
Conference_Location :
Providence, RI
ISSN :
1063-6919
Print_ISBN :
978-1-4577-0394-2
Type :
conf
DOI :
10.1109/CVPR.2011.5995567
Filename :
5995567
Link To Document :
بازگشت