基于轮廓的多模态医学图像配准方法.doc

文档介绍：
基于轮廓的多模态医学图像配准方法
钱鹰,李孟,韦庆杰*
(重庆邮电大学图形图像与多媒体实验室,重庆 400065)
5
10
15
20
25
30
35
40
摘要:针对基于灰度互信息的多模态医学图像配准中稳定性差且配准精度不高的问题,本文
利用同一器官在不同模态中轮廓具备相似性这一空间特点,在互信息理论的基础上提出一种
轮廓互信息的图像配准方法。首先使用形态学方法对待配准的两幅图像 R、F 求得对应的轮
廓图像 R'、F',然后利用轮廓互信息进行配准。实验结果表明,本文方法能够稳定地并且
准确地实现多模态医学图像配准。
关键词:灰度互信息;多模态医学图像;图像配准;轮廓;形态学
中图分类号:
A Multi-modal Medical Image Registration method Based
on Contour
Qian Ying, Li Meng, Wei Qingjie
(Chongqing University of Posts and munications, Chongqing, 400065, )
Abstract: There are some problems in Multi-modal image registration method based on gray
mutual information, such as poor robustness and lower registration accuracy. So a multi-modal
medical image registration method based on contours is proposed in this paper, which is based on
mutual information and using the spatial information that the an in different modal
contours have similarities. First, get the outline map R'、F' of the R、F which is subject to
registration through morphological method, then using contour mutual information in image
registration. Experimental results show that the method proposed in this paper can realize
multi-modal medical image registration robustness and accurately.
Key words: gray mutual information; multi-modal medical image; image registration;contour;
morphological
0 引言
单一模态的图像在实际的临床中通常不能够为医生提供足够的信息,因此需要将不同的
模态图像进行融合,从而得到更加丰富的信息以便医生了解