【摘要】 应用基于CT和MR图像等值特征表面的配准算法对多模医学图像进行了配准研究。在CT、MR图像中提取等值特征表面 ,进行图像的几何对准 ,并对结果进行初步评估 ,同时对该算法的稳健性 ,搜索最近点策略和插值策略进行了研究。结果表明 :这种方法能够达到亚象素级的配准精度 ,是一种稳健、高精度、全自动的配准方法更多还原

【Abstract】 An automated registration method based on equivalent features is discussed in this article. The skin surface of the head, the outer brain surface of MR and the inner skull surface of CT are commonly used as reference surfaces for image registration. Preliminary evaluation result showed that Sub pixel registration accuracy can be achieved with this approach. We also investigated the robustness of the method, strategies of finding the closest points and 3 D interpolation. The algorithm we presented is a very robust and fully automated multi modality registration method with high accuracy.更多还原