Graph-cut-based normalized coupled-curve evolution model

Abstract

Active contour models based on a level set function, which is a useful and implicit image representation tool, can be conveniently implemented. These level set-based active contour models optimized by a gradient decent flow method commonly lead to the problems of the active curves evoluting slowly and the segmentation results becoming highly dependent on initial locations. We develop a method using normalized graph cut to minimize the energy of the coupled-curve evolution model, which is based on the idea of maximizing the separation of the mean intensities of the two regions inside and outside the active curves. Using the minimum cut method, experiments demonstrate the computation time of the proposed method is much less than the level set-based method.