Unsupervised Sequence Segmentation by a Mixture of Switching Variable Memory Markov Sources.
Yevgeny SeldinGill BejeranoNaftali TishbyPublished in: ICML (2001)
Keyphrases
- fully unsupervised
- text segmentation
- segmentation algorithm
- blind separation
- segmentation of textured images
- bayesian image segmentation
- level set
- markov chain
- segmentation method
- image segmentation
- segmentation of natural images
- texture segmentation
- image analysis
- unsupervised learning
- data sources
- information sources
- semi supervised
- fully automatic
- memory requirements
- region growing
- markov model
- sequential data
- source separation
- medical images
- segmentation accuracy
- image processing
- markov fields
- unsupervised manner
- energy function
- main memory
- background subtraction
- medical imaging