Brain tumor detection and segmentation in a CRF (conditional random fields) framework with pixel-pairwise affinity and superpixel-level features.
Wei WuAlbert Y. C. ChenLiang ZhaoJason J. CorsoPublished in: Int. J. Comput. Assist. Radiol. Surg. (2014)
Keyphrases
- conditional random fields
- pairwise
- crf model
- semantic segmentation
- sequence labeling
- superpixels
- pixel level
- image labeling
- brain tumors
- markov random field
- contextual features
- probabilistic model
- higher order
- random fields
- maximum entropy
- discriminative models
- segmentation method
- graphical models
- named entity recognition
- information extraction
- image segmentation
- feature vectors
- object segmentation
- tumor segmentation
- similarity measure
- hidden markov models
- multi class
- semi supervised
- magnetic resonance images
- generative model
- parameter learning
- mri images
- low grade gliomas
- structured prediction
- approximate inference
- deformable models
- input image
- image processing
- segmentation algorithm
- web page prediction
- image regions