Mostly-Unsupervised Statistical Segmentation of Japanese Kanji Sequences
Rie Kubota AndoLillian LeePublished in: CoRR (2002)
Keyphrases
- segmentation of textured images
- japanese language
- native speakers
- text segmentation
- image segmentation
- bayesian image segmentation
- texture segmentation
- data driven
- fully unsupervised
- segmentation algorithm
- hidden markov models
- gaussian markov random fields
- statistical measures
- level set
- segmentation of natural images
- segmentation method
- sequential patterns
- region growing
- multiscale
- chinese characters
- supervised learning
- grey level
- shape prior
- statistical models
- multiresolution
- image analysis
- normalized cut
- unsupervised clustering
- agglomerative clustering
- energy function
- intra subject
- active contours
- medical images