Color Image Segmentation Based On The Fuzzy Support Vector Machine

Image segmentation is a technique and a process which can change the image into a feature with each area and it is the key technique of modern image analysis. But there is no universal algorithm for all color image segmentation, and most of these methods is extended from the gray scale image processing, which is not suitable for color image segmentation.Fuzzy technology is based on the fuzzy math theory. This paper is under the color space of HSI, and then uses the fuzzy support vector machine to cut up the color image.This paper started with the subject of color image segmentation which introduced the traditional color image segmentation technology and the traditional of image feature extraction, and then the main work was done as follows: the first, in chapter four introduced a fuzzy support vector machine based on gamma membership. The paper does some experiments in the HSI color space using the color features with color images for color image segmentation. Which compared to the traditional method had a good result; the second, in chapter five introduced another FSVM based on the linear combination kernel and KNN membership. The paper also do some experiments under the same conditions, and compared with the method previously that obtained more well results for segmentation and improved the segmentation performance.