HOGMep: Variational Bayes and higher-order graphical models applied to joint image recovery and segmentation.
Aurélie PirayreYuling ZhengLaurent DuvalJean-Christophe PesquetPublished in: ICIP (2017)
Keyphrases
- graphical models
- image recovery
- higher order
- belief propagation
- conditional random fields
- exponential family
- probabilistic model
- approximate inference
- probabilistic inference
- bayesian networks
- random variables
- structure learning
- image segmentation
- markov random field
- pairwise
- free energy
- level set
- log likelihood
- generative model
- latent variables
- markov chain monte carlo
- unsupervised learning
- message passing
- bayesian framework
- machine learning
- graph cuts
- information extraction
- support vector
- computer vision