Labeling colorectal NBI zoom-videoendoscope image sequences with MRF and SVM.
Tsubasa HirakawaToru TamakiBisser RaytchevKazufumi KanedaTetsushi KoideShigeta YoshidaYoko KominamiTaiji MatsuoRie MiyakiShinji TanakaPublished in: EMBC (2013)
Keyphrases
- image sequences
- scene labeling
- markov random field
- outdoor scenes
- hybrid method
- support vector machine svm
- support vector
- image segmentation
- support vector machine
- graph cuts
- feature selection
- mrf optimization
- motion analysis
- svm classifier
- spatio temporal
- knn
- generalization ability
- energy function
- multi class
- three dimensional
- motion estimation
- active learning
- narrow band
- support vectors
- video sequences
- higher order
- computer vision
- labeling problems
- sample selection
- scene understanding
- belief propagation
- high order
- object tracking
- kernel function
- optical flow
- classification algorithm
- svm classification
- feature vectors
- mrf model
- image frames
- training data
- low level vision
- image labeling
- camera motion
- neural network
- random fields
- feature space
- background subtraction
- structure from motion
- learning algorithm
- text classification
- level set
- parameter estimation
- text categorization
- max margin
- energy minimization
- training set
- conditional random fields