Gentle ICM energy minimization for Markov random fields with smoothness-based priors.
Zoran ZivkovicPublished in: J. Real Time Image Process. (2016)
Keyphrases
- energy minimization
- markov random field
- maximum a posteriori
- higher order cliques
- smoothness prior
- energy function
- graph cuts
- map estimation
- belief propagation
- mrf model
- bayesian framework
- smoothness constraint
- image segmentation
- random fields
- higher order
- spatial coherence
- parameter estimation
- image restoration
- labeling problems
- cost function
- prior information
- dual decomposition
- max flow
- pairwise
- prior distribution
- conditional random fields
- map inference
- posterior distribution
- low level vision
- potential functions
- loopy belief propagation
- interactive segmentation
- discontinuity preserving
- message passing
- iterated conditional modes
- mrf models
- image prior
- min cut
- post processing
- approximate inference
- potts model
- machine learning