Fast, Exact and Multi-scale Inference for Semantic Image Segmentation with Deep Gaussian CRFs.
Siddhartha ChandraIasonas KokkinosPublished in: ECCV (7) (2016)
Keyphrases
- multiscale
- image segmentation
- conditional random fields
- markov random field
- superpixels
- structured prediction
- loopy belief propagation
- crf model
- scale space
- random fields
- higher order
- max margin learning
- natural images
- image parsing
- graph cuts
- parameter learning
- image representation
- multiple scales
- image processing
- semantic segmentation
- fully connected
- probabilistic inference
- semantic similarity
- region growing
- maximum likelihood
- exact inference
- information extraction
- efficient learning
- gaussian mixture model
- graphical models
- maximum entropy
- wavelet transform
- natural language
- gaussian distribution
- computer vision
- boundary detection
- unsupervised image segmentation
- pairwise
- gray level
- image analysis
- hidden markov models
- edge detection
- generative model
- machine learning
- belief nets
- bayesian networks
- sequence labeling
- multiresolution
- belief networks
- active contours
- segmentation algorithm
- energy function
- zero crossing
- deformable models
- energy minimization
- object segmentation