改进的二阶互信息医学图像配准算法的研究  

Study on Medical Image Registration Based on Improved Genetic Algorithm and Improved Second-Order Mutual Information

作　　者：谭光兵[1] 叶汉民[1] 吴松华[1] 

机构地区：[1]桂林理工大学机械与控制工程学院,广西桂林541004

出　　处：《软件导刊》2011年第5期83-86,共4页Software Guide

基　　金：广西研究生教育创新计划项目(2009105960811M23)

摘　　要：针对人脑的二维图像设计了一种改进的遗传算法和二阶互信息相结合的医学图像配准算法,采用互信息配准模型,以图像的灰度统计信息为配准依据,采用改进的遗传算法搜索图像的最优变换参数,并比较一阶互信息配准与改进的二阶互信息配准两种算法,实验表明改进的二阶互信息配准方法具有较高的配准精度和稳定性。According to the 2D brain image,this paper designed a kind of medical image registration based on improved genetic algorithm and second-order mutual information,it took the mutual information as registration models,and used the gray-scale statistical information of images as registration basis,it used the improved genetic algorithm to search for the optimal transform parameter of images,and then compared first-order mutual information registration and second order mutual information registration,and the result showed that second-order mutual information registration method had higher registration accuracy and stability.

关 键 词：图像配准 遗传算法 一阶互信息 二阶互信息 

分 类 号：TP312[自动化与计算机技术—计算机软件与理论]