Non-supervised segmentation using multi-level Markov random fields.
Robert AzencottC. GraffignePublished in: ICPR (3) (1992)
Keyphrases
- markov random field
- image segmentation
- energy function
- mrf model
- low level vision
- textured images
- iterated conditional modes
- graph cuts
- mrf models
- label field
- energy minimization
- maximum a posteriori
- higher order
- belief propagation
- random fields
- markov random field model
- pairwise
- map estimation
- segmentation algorithm
- image restoration
- prior model
- min cut
- level set
- scene labeling
- potential functions
- parameter estimation
- object segmentation
- discriminative random fields
- labeling problems
- image analysis
- conditional random fields
- segmentation method
- message passing
- unsupervised segmentation
- loopy belief propagation
- multiscale
- iterative conditional
- active contours
- shape prior
- map inference
- potts model
- shape model
- texture segmentation
- unsupervised image segmentation
- non stationary
- smoothness prior
- medical images
- efficient inference
- gibbs energy