SAR image segmentation through information-theoretic heterogeneity features and tree-structured Markov random fields.
Bruno AiazziStefano BarontiLuciano AlparoneGiovanni CuozzoCiro D'EliaGilda SchirinziPublished in: IGARSS (2005)
Keyphrases
- information theoretic
- markov random field
- mutual information
- information theory
- belief propagation
- energy minimization
- random fields
- image segmentation
- potential functions
- maximum a posteriori
- pairwise
- graph cuts
- min cut
- energy function
- parameter estimation
- graphical models
- higher order
- image features
- markov networks
- kullback leibler divergence
- feature vectors
- feature selection
- sar image segmentation
- kl divergence
- gaussian mixture model