利用高维互信息的多模态医学图像配准

Multi-modality medical image registration using high dimension mutual information

互信息相似性测度在多模态医学图像配准中获得了广泛的应用,然而其不足之处在于没用充分利用图像固有的空间信息。针对这一不足,提出了利用图像邻域信息的高维互信息配准方法。首先用图像像素及其邻域构成高维向量的集合,然后利用基于最近邻的熵估计法来估计集合的高维熵,并采用近似最近邻搜索算法来加快高维熵的计算。实验结果验证了新的相似性测度的有效性。

Mutual Information（MI） based similarity measure has been widely used in muhi-modal medical image registration.One drawback of MI, however, is that it fails to take into account the spatial information of image.In this paper,we propose a novel extension to MI called High-dimension Mutual Information（HMI）.This extension takes advantage of neighbor pixels to construct a set of high dimension vector,and then evaluates high dimension entropy using an algorithm based on Nearest Neighbor Search （NNS）.In order to compute HMI effectively,we adopt an algorithm called approximate nearest neighbor search.Experiments demonstrate the validation of the extended methods.