A deep quasi-linear kernel composition method for support vector machines.
Weite LiJinglu HuBenhui ChenPublished in: IJCNN (2016)
Keyphrases
- support vector
- classification accuracy
- cost function
- similarity measure
- high accuracy
- experimental evaluation
- support vector machine
- neural network
- computationally efficient
- synthetic data
- high precision
- dynamic programming
- significant improvement
- prior knowledge
- edge detection
- feature set
- segmentation algorithm
- detection algorithm
- training examples
- segmentation method
- kernel methods