On iterative source-channel image decoding with Markov random field source models.
Jörg KliewerNorbert GörtzAlfred MertinsPublished in: ICASSP (4) (2004)
Keyphrases
- markov random field
- random fields
- mrf models
- mrf model
- image segmentation
- energy function
- neighboring pixels
- prior model
- parameter estimation
- potential functions
- graph cuts
- texture model
- textured images
- belief propagation
- input image
- energy minimization
- map estimation
- maximum a posteriori
- single image
- conditional random fields
- spatial continuity
- multiscale
- bayesian estimation
- markov random field model
- contextual constraints
- image restoration
- image pixels
- mrf optimization
- bayesian framework
- maximum a posterior
- long range
- segmentation algorithm
- higher order
- pairwise
- gauss markov random fields
- message passing
- maximum entropy
- test images
- maximum likelihood
- texture synthesis
- pixel values
- smoothness prior
- probabilistic model
- similarity measure