基于互信息和小波分解的混合优化医学图像配准

Medical Image Registration-Based on Methodologies of Mutual Information,Wavelet Decomposition,and Hybrid Optimization

提出一种新的基于互信息和小波分析的混合优化配准算法,在最低分辨率下用模拟退火算法进行全局优化,在高分辨率下采用Powell算法进行优化.结果显示:该方法用于医学图像配准,并与已有的传统方法进行分析比较,具有良好的全局优化性能和时间性能.

In this paper, a new method based on mutual information, wavelet decomposition, and hybrid optimization is proposed for medical image registration. Two kinds of optimization were performed. One is global optimization using simulate anneal algorithm under the lower resolution rate and the other is to employ the Poweli algorithm under higher resolution rate. Results show that a new method is suitable to medical image registration and has a good global optimization and runtime performance.