Bayesian contextual classification based on modified M-estimates and Markov random fields.
Yonhong JhungPhilip H. SwainPublished in: IEEE Trans. Geosci. Remote. Sens. (1996)
Keyphrases
- markov random field
- random fields
- energy minimization
- discriminative random fields
- graph cuts
- maximum a posteriori
- belief propagation
- pairwise
- higher order
- mrf model
- potential functions
- parameter estimation
- image segmentation
- energy function
- map inference
- efficient inference
- loopy belief propagation
- text classification
- machine learning
- conditional random fields
- maximum entropy
- message passing
- low level vision
- support vector
- map estimation
- potts model
- feature vectors
- feature extraction
- piecewise constant functions
- prior distribution
- textured images
- markov networks
- posterior distribution
- document classification
- closed form
- computer vision