Subspace based active contours with a joint distribution metric for semi-supervised natural image segmentation.
Shu-Juan PengXin LiuYiu-ming CheungPublished in: ICASSP (2012)
Keyphrases
- active contours
- joint distribution
- image segmentation
- semi supervised
- level set method
- level set
- active contour model
- maximum entropy
- object boundaries
- shape prior
- random variables
- geodesic active contours
- texture image segmentation
- energy function
- conditional probabilities
- curve evolution
- segmentation method
- semi supervised learning
- gradient vector flow
- energy functional
- labeled data
- supervised learning
- probability distribution
- unlabeled data
- computer vision
- gvf snake
- image processing
- pairwise
- principal component analysis
- region growing
- gradient flow
- graph cuts
- energy minimization
- segmentation algorithm
- gradient flows
- multiscale
- deformable models
- active learning
- feature space
- covariance matrices
- learning algorithm
- markov random field
- force field
- low dimensional
- image segmentation methods
- external force
- geometric active contours
- high dimensional
- gaussian distribution