A Geometrical Method to Improve Performance of the Support Vector Machine.
Peter WilliamsSheng LiJianfeng FengSi WuPublished in: IEEE Trans. Neural Networks (2007)
Keyphrases
- support vector machine
- significant improvement
- high accuracy
- detection method
- preprocessing
- classification method
- objective function
- computational complexity
- fully automatic
- high precision
- segmentation method
- optimization algorithm
- segmentation algorithm
- computationally efficient
- input data
- dynamic programming
- cost function
- support vector
- training data
- denoising
- experimental evaluation
- feature set
- image classification
- prior knowledge
- evolutionary algorithm
- synthetic data
- similarity measure
- image processing
- feature selection