Iteratively reweighted graph cut for multi-label MRFs with non-convex priors.
Thalaiyasingam AjanthanRichard I. HartleyMathieu SalzmannHongdong LiPublished in: CVPR (2015)
Keyphrases
- multi label
- labeling problems
- graph cuts
- markov random field
- convex relaxation
- multi label classification
- energy minimization
- globally optimal
- belief propagation
- energy function
- shape prior
- object segmentation
- binary classification
- spatial coherence
- image segmentation
- max flow
- image annotation
- segmentation algorithm
- text categorization
- text classification
- interactive segmentation
- higher order
- class labels
- image classification
- approximate inference
- maximum likelihood
- loopy belief propagation
- min cut
- multiple labels
- feature selection
- potts model