Unsupervised multi-class segmentation of SAR images using triplet Markov fields models based on edge penalty.
Yan WuMing LiPeng ZhangHaitao ZongPing XiaoChunyan LiuPublished in: Pattern Recognit. Lett. (2011)
Keyphrases
- multi class
- markov fields
- unsupervised image segmentation
- sar images
- random fields
- parameter estimation
- small samples
- support vector machine
- texture segmentation
- edge detector
- level set
- feature selection
- image analysis
- pairwise
- edge detection
- maximum likelihood
- model selection
- probabilistic model
- image segmentation
- textured images
- computer vision
- knn
- markov property
- unsupervised segmentation
- training data
- multiscale
- sample size
- non stationary
- machine learning
- object detection
- semi supervised