Computer Engineering and Applications ›› 2009, Vol. 45 ›› Issue (10): 154-156.DOI: 10.3778/j.issn.1002-8331.2009.10.046

• 图形、图像、模式识别 • Previous Articles     Next Articles

Application of mutual information of regional in HEAD MRI registration

LI Xiao-guang,TAN Jian-hao   

  1. College of Electrical and Information Engineering,Hunan University,Changsha 410012,China
  • Received:2007-08-21 Revised:2007-11-16 Online:2009-04-01 Published:2009-04-01
  • Contact: LI Xiao-guang

邻域互信息在磁共振颅脑图像配准的应用

李晓光,谭建豪   

  1. 湖南大学 电气与信息工程学院,长沙 410012
  • 通讯作者: 李晓光

Abstract: This paper proposes a novel extension to mutual information called regional mutual information.In the process of computing the image information entropy,this method not only considers the information of each pixel,but also considers the neighborhood of the pixel,thus increases the stable of registration,reduces the local optimization according by linear interpolation algorithm and partial volume interpolation,and effectively smoothes the curve of information entropy.Then uses the mixed optimization algorithm based on GA algorithm and Powell algorithm to optimize the function of information entropy.The results show that the modified MRI measure is very accuracy and effictive.

摘要: 引入了一种新的对互信息的扩展称之为邻域互信息。在计算图像信息熵的过程中,不仅考虑了图像中每点像素的信息,还考虑了每点像素邻域像素的信息,从而增加了配准过程中的稳定性。并抑制了配准过程中,因线性插值和部分体积插值所产生的局部极值情况,平滑了信息熵的曲线。然后采用基于遗传算法和Powell所组成的混合优化算法对信息熵函数进行寻优。实验结果表明该算法具有良好的精确性和有效性。