Abstract: Image segmentation is an important research field of pattern recognition, image understanding and computer vision. Support vector machine (SVM) is now widely used in image segmentation, but the training samples are usually selected artificially. hhis will reduce the self-adaptability and affect the classification performance of image segmenta- lion. This paper presented a color image segmentation SVM approach based on training samples automatic selection. First,Fuzzy C-Means (FCM) clustering algorithm was used to obtain the training samples for SVM automatically. hhen, color and texture features were extracted from the image as attributes of training samples of SVM. Finally, the images were segmented by the trained classifier. The experiment results demonstrate that the proposed approach can a- chieve good segmentation performance.

Key words: Image segmentation,Support vector machine,Fuzzy C-means