A Fast Hybrid K-Means Level Set Algorithm for Image Segmentation
Qiuju Wei · Zhongguo yixue wulixue zazhi · 2011
Objective:To increase the speed of level set image segmentation algorithm.Methods:A fast hybrid K-Means level set method is proposed for image segmentation in this paper.We firstly draw a connection between a level set algorithm based on Mumford-Shah model and K-Means plus nonlinear diffusion preprocessing,analysing their drawbacks.Then,we exploit this link to develop a new hybrid numerical technique for segmentation that draws on the speed and simplicity of K-Means procedures,and the robustness of level set algorithms.Results:The proposed method costs lesser time than standard level set methods.Conclusions:Taking advantage of K-Means simplicity and efficientibility,this new level set algorihm increases the speed of image segmentation.So it is practical.