Chinese Journal of Computers   Full Text
Title	Medical Image Registration Using Co-Occurrence Mutual Information
Authors	LU Zhen-Tai CHEN Wu-Fan
Address	(Key Laboratory for Medical Imaging of Southern Medical University, Guangzhou 510515)
Year	2007
Issue	No.6(1022—1027)
Abstract &   Background	Abstract Mutual Information(MI) is calculated on a pixel by pixel basis. It takes into account only the relationship between corresponding individual pixels and not those of each pixel’s respective neighborhood. It ignores the spatial information. This paper proposes a new measure—Co-occurrence mutual information(Co-MI). It is an extension to the mutual information framework which incorporates spatial information about image structure into the registration process and has the potential to improve the accuracy and robustness of image registration. The results indicate that co-occurrence mutual information is a more robust similarity measure for image registration than MI. It is more important that Co-MI can be used in other scope, such as economics, operational research and pattern recognition, just like MI. keywords image registration; mutual information; co-occurrence matrix; co-occurrence mutual information; spatial information background This work is supported in part by the National Basic Research Program(973 Program) of China(No.2003CB716103). In this project, medical images and electrophysiological signals are the main research focuses, which are also key problems in the current field of medical information processing. The research outcomes of the project not only enriches the content of life science and information science, and promote the development of these fields,but also become a kind of new knowledge economy in post-processing software of medical clinical information. Especially, it will provide new intellectual properties for independently development of medical imaging equipments in China. The research group’s interests include image understanding and analysis, image segmentation, registration and medical image computing. The group has proposed some novel algorithms about image registration, such as elastic registration algorithm of medical image based on the apriori knowledge and Markov Random Field model, mixing rigid-elastic multiresolution algorithm for medical image registration, etc. And a new registration method based on co-occurrence mutual information is presented in the paper.