Recently, a family of global, non-submodular energy functions has been proposed that is expressed as coupling edges in a graph cut. This formulation provides a rich modelling framework and also leads to efficient approximate inference algorithms. So far, the results addressed binary random variables. Here, we extend these results to the multi-label case, and combine edge coupling with move-making algorithms.
| Author(s): | Jegelka, S. and Bilmes, J. |
| Pages: | 1-4 |
| Year: | 2011 |
| Month: | June |
| Day: | 0 |
| Bibtex Type: | Conference Paper (inproceedings) |
| Event Name: | CVPR 2011 Workshop on Inference in Graphical Models with Structured Potentials |
| Event Place: | Colorado Springs, CO, USA |
| Digital: | 0 |
| Electronic Archiving: | grant_archive |
| Links: | |
BibTex
@inproceedings{JegelkaB2011_4,
title = {Multi-label cooperative cuts},
abstract = {Recently, a family of global, non-submodular energy functions
has been proposed that is expressed as coupling edges
in a graph cut. This formulation provides a rich modelling
framework and also leads to efficient approximate inference
algorithms. So far, the results addressed binary random variables. Here, we extend these results to the multi-label case, and combine edge coupling with move-making algorithms.},
pages = {1-4},
month = jun,
year = {2011},
slug = {jegelkab2011_4},
author = {Jegelka, S. and Bilmes, J.},
month_numeric = {6}
}