Human brain image registration using mutual information
Yang Hu · Zhongguo yixue wulixue zazhi · 2001
There has recently been considerable interest in the medical image processing community,in the use of mutal informa -tion for automatic registration of images from different modalities.Applying registration methods to multimodality medical images is beneficail in neurology to study structure-function relationships and guide neurosurgery.This paper describes an automated mu tu-al information-based method for human brain image registration.We have applied this approach to the problem of geometric alignment of primate autoradiographs.Preliminary evaluation results are presented below on our experiments.We also investigated the performance of normalized MI,multiresolution strategies,various interpolation and optimization methods for maximization of mu -tual information,aiming at increasing registration speed and accuracy.Because no assumptions were made regarding the nuture of the relation between the image intensities in both modalities,mutual information-based method is very robust and may be quite useful in multimodality voxel-based registration.