A Riemannian Stopping Criterion for Unsupervised Phonetic Segmentation.
Ciro Gracia PonsXavier AngueraXavier BinefaPublished in: ICMLA (2) (2013)
Keyphrases
- stopping criterion
- region merging
- segmentation algorithm
- convergence rate
- fully unsupervised
- region growing
- image segmentation
- machine learning
- multiscale
- level set
- image analysis
- medical images
- higher order
- speech recognition
- semi supervised
- cost function
- test images
- merging algorithm
- multiple objects
- shape prior
- deformable models
- segmentation method
- clustering algorithm