Restoration and Segmentation of Highly Degraded Characters Using a Shape-Independent Level Set Approach and Multi-level Classifiers.
Reza Farrahi MoghaddamDavid Rivest-HénaultMohamed CherietPublished in: ICDAR (2009)
Keyphrases
- region based segmentation
- level set
- shape prior
- image segmentation
- active contour model
- chan vese
- segmentation algorithm
- graph cuts
- prior shape
- energy functional
- multiscale
- object boundaries
- variational framework
- segmentation method
- image restoration
- perceptual grouping
- level set segmentation
- texture segmentation
- training data
- active appearance models
- prior information
- support vector
- chain code
- energy minimization
- region growing
- degraded images
- shape features
- medical images
- training set
- prostate segmentation
- object shape
- discriminative classifiers
- image processing
- shape representation
- feature selection
- decision trees
- partial differential equations
- deformable models
- energy function
- segmented images
- shape variations
- shape model
- geometric shapes
- maximum likelihood criterion
- level set method
- video sequences
- feature set
- active contours
- global shape
- shape descriptors
- statistical shape model
- curve evolution