A convex approximation of regularization models for motion estimation with Markov random fields.
Mustapha KardouchiAlbert DipandaFranck MarzaniLouis LegrandPublished in: ICIP (1) (1996)
Keyphrases
- markov random field
- potential functions
- random fields
- motion estimation
- parameter estimation
- maximum a posteriori
- image segmentation
- graph cuts
- higher order
- conditional random fields
- partition function
- efficient inference
- energy minimization
- image sequences
- belief propagation
- image restoration
- energy function
- mrf models
- prior models
- map inference
- mrf model
- undirected graphical models
- low level vision
- potts model
- textured images
- markov networks
- message passing
- prior information
- graphical models
- optical flow
- pairwise
- video sequences