Algorithm of Texture Segmentation Combining FCM and FSVM

Fuzzy support vector machine(FSVM), as a improved support vector machine(SVM), is a powerful classifier in texture image segmentation because of its outstanding performance in high-dimensional classification. However, FSVM, as a supervised classification algorithm, is unable to classify the features of the texture image automatically because training data are needed. To solve this problem, a texture segmentation algorithm combing the fuzzy C-means clustering algorithm(FCM) and FSVM is proposed. The FCM provides training data and automatic texture segmentation is carried out. Experiments based on Matlab 7.0 are carried through, and this algorithm has presented its validity in improving segmentation precision.