Multiscale Markov random fields and constrained relaxation in low level image analysis.
Patrick PérezFabrice HeitzPublished in: ICASSP (1992)
Keyphrases
- markov random field
- image segmentation
- multiscale
- image analysis
- low level
- scale space
- image processing
- mrf model
- high level
- graph cuts
- low level vision
- parameter estimation
- belief propagation
- figure ground
- random fields
- energy function
- image understanding
- higher order
- energy minimization
- pairwise
- maximum a posteriori
- potts model
- image restoration
- computer vision
- conditional random fields
- image representation
- natural images
- potential functions
- wavelet transform
- map inference
- edge detection
- active contours
- message passing
- iterative conditional
- piecewise constant functions
- loopy belief propagation
- efficient inference
- level set
- expectation maximization
- texture analysis