Improving foreground segmentations with probabilistic superpixel Markov random fields.
Alexander SchickMartin BäumlRainer StiefelhagenPublished in: CVPR Workshops (2012)
Keyphrases
- markov random field
- image segmentation
- graph cuts
- belief propagation
- random fields
- higher order
- energy minimization
- mrf model
- maximum a posteriori
- parameter estimation
- image restoration
- object segmentation
- energy function
- factor graphs
- textured images
- pairwise
- potts model
- foreground objects
- potential functions
- message passing
- superpixels
- image segmentation algorithm
- map estimation
- map inference
- low level vision
- efficient inference
- conditional random fields
- generative model
- medical images
- foreground and background
- probabilistic networks
- piecewise constant functions
- posterior probability
- background subtraction
- graphical models
- bayesian networks
- shape prior
- figure ground
- iterative conditional
- discriminative random fields
- loopy belief propagation
- approximate inference
- long range
- maximum entropy
- conditional probabilities
- segmentation algorithm
- denoising
- input image
- image sequences