摘要
本文提出一种新的基于轮廓提取和最大互信息理论的医学图像配准的算法。该算法克服了在配准过程中存在鲁棒性因素、出现误配情况及互信息单一的利用图像灰度信息的局限性,将边缘检测与互信息相结合,提高了原有算法的性能,可较准确地完成图像配准任务。本文并对提出的配准算法进行了Matlab仿真实验,对仿真结果进行分析。
A new algorithm of medical image registration in contour-based extraction and maximum mutual information was proposed.This algorithm could overcome the limitation in the use of mutual information of a single gray-scale image and improve the performance of the original algorithm,finish the task of image registration accurately by combining the edge detection and mutual information.Matlab simulation was also performed,and the results were analyzed for the registration algorithms in this paper.
出处
《中国介入影像与治疗学》
CSCD
2010年第3期325-327,共3页
Chinese Journal of Interventional Imaging and Therapy
关键词
医学图像配准
边缘检测
最大互信息
Medical image registration
Edge detection
Maximum mutual information
