A Context-Sensitive Clustering Technique Based on Graph-Cut Initialization and Expectation-Maximization Algorithm.
Mayank TyagiFrancesca BovoloAnkit K. MehraSubhasis ChaudhuriLorenzo BruzzonePublished in: IEEE Geosci. Remote. Sens. Lett. (2008)
Keyphrases
- expectation maximization
- k means
- context sensitive
- globally optimal
- graph cuts
- em algorithm
- learning algorithm
- energy function
- segmentation algorithm
- clustering algorithm
- hierarchical clustering
- clustering method
- interactive segmentation
- data clustering
- maximum likelihood
- probabilistic model
- mixture model
- cluster analysis
- semi supervised
- spectral clustering
- natural language
- similarity measure
- loopy belief propagation
- image segmentation