基于互信息的复合遗传算法在医学图像配准中的应用

Application of Complex Genetic Algorithm Based on Mutual Information to Medical Image Registration

分析当前的医学图像配准技术及算法,选择基于体素相似性的图像配准方法,以互信息作为配准相似性测度,将混合遗传算法、自适应遗传算法和育种算法相结合,提出复合遗传算法,以提高遗传算法的鲁棒性,加速收敛过程,准确地实现多模医学图像配准,再利用此算法对无噪声图像和有噪声图像进行配准实验.实验结果证明了该算法的可行性,是一种值得推广的医学图像配准算法.

The current medical image registration techniques and algorithms were analyzed to select image registration method on the basis of voxel similarity.Mutual information as measure of registration similarity,hybrid genetic algorithm,adaptive genetic algorithm and breeding algorithm were combined to propose complex genetic algorithm to improve the robustness of the genetic algorithm,to speed up the convergence process and accurately realize multi-mode medical image registration.This algorithm was re-used to do noise-free image and noisy images registration.Experimental results prove the feasibility of the method,with some theoretical significance and high practical value,which is deserved to be promoted.