A Medical Image Segmentation Based on Global Variational Level Set

摘要

A medical image segmentation based on global variables differential level set is proposed in this paper for medical images with complex topological structure, strong contrast and low noise characteristics. It make full use of the image area information, build a energy model, and using variation gradient information to establish a global energy model to get the minimization value, which is geodesic active contour (GAC) model. Experimental results show that the method set in the initial outline of the evolution without success to avoid the reinitialization and correction process, thus saving computing time. With traditional methods and TV and CV method, the method convergence stable segmentation accuracy is good, easy parameter adjustment and split speed, better medical treatment of low contrast, blurred image.