A kernel induced energy based active contour method for image segmentation

Active contour model is a promising method in image segmentation. However, existing active contour model and its evolution often suffer from slower convergence rates and easily to be trapped in local optima due to the presence of noise. In this paper, a novel curve evolution model based on kernel mapping method is presented. The method first transforms original image data into a kernel-induced space by a kernel function. In the kernel-induced space, the kernel-induced non-Euclidean distance between the observations and the regions parameters is integrated to formulate a new level set based active contour model. The method proposed in this paper leads to a flexible and effective alternative to complex model the image data. In the end of this paper, detailed experiments are given to show the effectiveness of the method in comparison with conventional active contour model methods.