一种基于广义互信息的图像配准算法

Image registration algorithm based on generalized mutual information

互信息作为图像配准中的相关度矩阵有着广泛的应用,通常采用的是基于Shannon熵的互信息。采用一个广义的信息熵——Renyi熵,提出了一种基于广义互信息的图像配准方法。在全局搜索阶段,采用q取较小值的Renyi熵,此时,Renyi熵可以消除局部极值,再通过局部优化方法对当前的局部最优解进行局部寻优,以找到全局最优解;在局部优化阶段,使用基于q→1时的Renyi熵的归一化互信息测度作为目标函数。实验结果表明:相对于归一化互信息图像配准算法,基于Renyi熵的互信息配准算法有良好的配准效果,且提高了配准速度。

Mutual information is widely used as a similarity metric for image registration. Usually similarity metric is based on Shannon definition of entropy. One generalized entropy named as Renyi entropy is used and a new image registration algorithm is. proposed based on generalized mutual information. Firstly Renyi entropy with smaller q is used for global searches, here Renyi entropy can remove some unwanted local optimum ; secondly, the local one is used to locate the global optimal solution by searching the current local optimal ones. The normalized mutual information measure based on Renyi entropy with q→1 is taken as the objective function. The experiment results show that the mutual information image registration algorithm based on Renyi entropy is better than the normalized mutual informatron image registration algorithm, and the speed of registration is improved.