基于变异系数的分片常值图像快速分割  

FAST SEGMENTATION OF IMAGES WITH CONSTANT PIECEWISE INTENSITIES BASED ON COEFFICIENT OF VARIATION

作　　者：刘景霞[1] 何传江[1] 

机构地区：[1]重庆大学数学与统计学院,重庆401331

出　　处：《计算机应用与软件》2014年第11期197-200,共4页Computer Applications and Software

基　　金：重庆市自然科学基金计划项目(CSTC;2010BB9218)

摘　　要：结合C-V模型和变异系数,提出一种水平集演化方程为常微分方程类型的模型。该模型不仅能正确分割分片常值图像,而且迭代次数不受初始轮廓大小、位置和形状的影响,且水平集函数无需重新初始化,从而能够快速分割分片常值图像,同时又能分割C-V模型、ICV模型和PSM模型不能正确分割的3-phases图像。In this paper we propose a model to construct the level set evolution equation as an ODE（ ordinary differential equation） in combination with C-V model and coefficient of variation.The model can correctly segment the image with constant piecewise intensity,and the iteration times will not be affected by the size,position and shape of the initial contour.Furthermore,the level set function does not need to be reinitialised,so that it can fast segment the image with constant piecewise intensity,and can segment those 3-phases images such as C-V model,ICV model and PSM model which are difficult to be segmented correctly at the same time.

关 键 词：变异系数 C-V模型 水平集函数 主动轮廓 PSM模型 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]