Research And Application Of Active Contour Model In Image Segmentation Algrithm

Image segmentation is one of basic research issues of image process and computer vision,which is is used to extract the target of interest from the image. In recent years,more and more research is focused on the image segmentation, and many new theories and methods are proposed. However there is still no universal method of image segmentation. The edge of the image is an very important image feature and available information. Through capturing the edge information of the target, the traditional active contour model, referred to as ACM, also known as Snake model, gives a unified solution approach to a series of computer vision problem. Active contour model(ACM) give a unified solution on a range of Computer Vision issues by using the information of the target edge. So it has been already applied in many areas of Computer Vision,for example:the edge extraction, image segmentation and classification, motion tracking, three-dimensional reconstruction, stereo matching, and so on.The paper is widely applying in the existed active contour models, the active contour model based on particle swarm optimization, referred to PSO, is proposed. Using the simple and easy implement, fast convergence and the large probability of particle swarm optimization algorithm to find the global optimal solution of optimization problems, the fast and accurate target edge segmentation can be achieved.The following is done in thesis:1,Based on the standard particle swarm optimization algorithm, an improved Multi-Particle Swarm Co-evolution Algorithm is presented. Through the co-evolution between adjacent particles and periodically updating the share information,it have increased the computational speed of the algorithm, and improved the solving performance.2,This paper presented the improved particle swarm optimization algorithm based on the active contour model, it expands the search region of control points ,and then achieves an accurate segmentation of the target border. Compared with the traditional active contour model, by the co-evolution between adjacent particles,it can search the depression border of the target more accurately without additional time.