New Kernel Methods For Image Segmentation

How to separate a target from the background?That's a trouble in machine vision systems, and the key to solve this problem is image segmentation. The results of image segmentation will directly affect the performance of vision system, which is about image analysis and understanding. While the computer technology developing rapidly, the research of image segmentation has always been one of the central issues for researchers in the field of image process, which results in an amount of algorithms for image segmentation. And these algorithms have been widely used. in areas like computer vision, pattern recognition, medical image processing, etc.So far, although many algorithms for image segmentation have emerged, there is no general one .Most of the algorithms is intended for some specific images. In the research of algorithms for image segmentation, the problem that often needs most consideration is the suppression against the numerous interference noises in real images and the improvement of calculation speed,This paper presents an improved support vector machines for image segmentation. This algorithm not only inherits the general ability of support vector machine classification, generalization ability and other characteristics; but also has incremental learning, fast learning , less support vectors and so on,. And it can be as twice faster as LSSVM learning speed,Finally, this paper summarizes a lot of results about image segmentation in these years, and suggests an improved SVM for image segmentation. It's very useful for image segmentation. The incremental learning features and fast performance make it a better segmentation of modern image of the growing demand.