基于改进的C-V模型虚拟人脑图像分割模型  

Chinese visual human images segmentation based on improved C-V model

作　　者：陈允杰[1,2] 张建伟[1] 王利[2] 王平安[3] 夏德深[2] 

机构地区：[1]南京信息工程大学数学系,南京210044 [2]南京理工大学计算机科学与技术学院,南京210094 [3]香港中文大学计算机科学与工程系,中国香港沙田

出　　处：《计算机工程与应用》2008年第16期13-17,22,共6页Computer Engineering and Applications

基　　金：国家自然科学基金(the National Natural Science Foundation of China under Grant No.60773172);香港特区政府研究资助局研究项目(No.CUHK/4185/00E);香港中文大学研究基金(No.2050345);江苏省青蓝工程;南京信息工程大学科研基金

摘　　要：C-V模型是一种较为经典的分割模型,但传统的C-V模型仅能够将图像分割成单一的目标部分与背景部分;用于彩色图像分割往往基于目标的强度信息;在曲线演化过程中需要重新初始化水平集函数保持符号距离函数。针对这些问题,使用PCA理论将颜色空间投影到新的空间中,可以扩大两者的颜色距离;使用局部信息可校正颜色强度不均匀;将距离约束项引入到模型中,使模型能够无需重新初始化,提高了演化速度。实验结果表明改进的算法能较精确地得到分割结果。C-V model is one of the best segmentation methods,but the classical C-V models only segment the image into object and background;only use the intensity information when segmenting color images;must re-initial the distance function during evolving the curves.In Chinese Visible Human（CVH） images,there are many fake grey matters and with the effects of these fake matters the C-V model can hardly separate grey matters with fake grey matters.To deal with the problem the PCA model is presented to large the difference of grey matters and fake grey matters.With the effects of tissues themselves,there are many inhomogenous phenomenons in the CVH images;the local information is added to model to reduce these effects.Using the distance resistance energy,the model can evolve curves without re-initialization.The Chinese visual human brain images segmentation experimental results show that the method of this paper can get fight results in an accuracy way.

关 键 词：中国虚拟人 C—V模型 主成分分析 局部信息 距离约束项 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]