Fully connected CRF with data-driven prior for multi-class brain tumor segmentation.
Haocheng ShenJianguo ZhangPublished in: ICIP (2017)
Keyphrases
- fully connected
- multi class
- conditional random fields
- brain tumor segmentation
- pairwise
- support vector machine
- multi class classification
- prior knowledge
- multiple classes
- error correcting output codes
- magnetic resonance images
- feature selection
- scale free
- graphical models
- multi class classifier
- higher order
- markov random field
- multi class boosting
- segmentation method
- artificial neural networks
- binary classifiers
- data sets