Multi-scale model and level set based image segmentation method and its medical application

Abstract

To overcome the deficiency of conventional methods in image segmentation influenced by image noise, level set image segmentation based on Mumford-Shah model is often used. But it has the drawback of large calculation. This paper proposes a multi-scale algorithm based on Mumford-Shah method. Experimental results show that the given multi-scale algorithm can greatly decrease the amount of computation and enhance the noise suppressing ability. Finally the algorithm is successfully applied to segmentation of images of abdominal macrophages of mice. © 2007 IEEE.