| Image segmentation has been widely applied in many fields. Up to now, there is not a common segmentation method for any type of images. In this paper, we mainly focus on the active contours methods, from the principle, algorithm to experiments.In this thesis, we propose a new algorithm based on the Chan-Vese model (CV) that is very efficient in image segmentation. However, when the original image is contaminated by noise, the standard level set method for solving the CV model is very sensitive to the initial level set function and the parameter of the length of the evolving contour. Here, we propose a two-step segmentation algorithm, where, in the first step, a coarse segmentation is obtained by using some traditional method, and in the second step, the coarse segmentation is used as an initial solution in the CV model to find a better segmentation. Moreover, we give a model for adaptively selecting the parameter of the length of the evolving contour, where the parameter is defined as an increasing function of the noise variance. The two-step segmentation method with adaptive parameter selection ensures not only automatic evolution but also fast and accurate partition.Experiment on computer-generated images and real images shows that the algorithm proposed here is very efficient. |
