The Research Of Level Set Method In Image Segmentation

In the research and application of image, image segmentation is an important part of image processing. People often only interested in certain parts of the image, according to different image analysis task, image segmentation aims at partition the original image into several meaningful regions or isolating the region of interesting.In this paper, several traditional image segmentation methods are introduced firstly, then the basic theory on level set method is systematically researched. The level set method based on geometric deformable model, which translates the problem of evolution of 2-D close curve into the evolution of level set function in the space with higher dimension. Geometric deformable model is made the widespread application in image segmentation because of auto-adapted ability of topology and the steady value condition. And Geodesic Active Contour Model and Momford-Shah Model as well as their application in image segmentation field is discussed.This article prime task is as follows: firstly the GAC model and simply M-S model are programmed, and the drawbacks of these methods is pointed out through experiments analysis. The improved method is proposed which combines GAC model and simply M-S model(C-V method), the method consider the information of gradient information as well as homogeneity region information. The experimental results show that this method can make up the defects of them. Using this method, the overall solution can be received, and the complexity of running time can be reduced. In addition C-V double level set method is proposed, It has an ideal scheme of region partition and is able to segment the image which has many different regions.