Restoration of noisy regions modeled by noncausal Markov random fields of unknown parameters.
Amir WaksOleh J. TretiakGeorge K. GregoriouPublished in: ICPR (2) (1990)
Keyphrases
- markov random field
- image restoration
- neighboring pixels
- random fields
- higher order
- parameter estimation
- degraded images
- maximum a posteriori
- mrf model
- belief propagation
- graph cuts
- image segmentation
- energy function
- pairwise
- energy minimization
- piecewise constant functions
- loopy belief propagation
- conditional random fields
- iterative conditional
- potential functions
- map estimation
- input image
- textured images
- image deblurring
- potts model
- low level vision
- partition function
- image features
- image processing
- information extraction
- efficient inference
- texture model
- edge detection
- message passing
- expectation maximization