Generalized Fast Approximate Energy Minimization via Graph Cuts: A-Expansion B-Shrink Moves

Abstract

We present α-expansion β-shrink moves, a simple generalization of the widely-used αβ-swap and α-expansion algorithms for approximate energy minimization. We show that in a certain sense, these moves dominate both αβ-swap and α-expansion moves, but unlike previous generalizations the new moves require no additional assumptions and are still solvable in polynomial-time. We show promising experimental results with the new moves, which we believe could be used in any context where α-expansions are currently employed.

Cite

Text

Schmidt and Alahari. "Generalized Fast Approximate Energy Minimization via Graph Cuts: A-Expansion B-Shrink Moves." Conference on Uncertainty in Artificial Intelligence, 2011.

Markdown

[Schmidt and Alahari. "Generalized Fast Approximate Energy Minimization via Graph Cuts: A-Expansion B-Shrink Moves." Conference on Uncertainty in Artificial Intelligence, 2011.](https://mlanthology.org/uai/2011/schmidt2011uai-generalized/)

BibTeX

@inproceedings{schmidt2011uai-generalized,
  title     = {{Generalized Fast Approximate Energy Minimization via Graph Cuts: A-Expansion B-Shrink Moves}},
  author    = {Schmidt, Mark and Alahari, Karteek},
  booktitle = {Conference on Uncertainty in Artificial Intelligence},
  year      = {2011},
  pages     = {653-660},
  url       = {https://mlanthology.org/uai/2011/schmidt2011uai-generalized/}
}