摘要
多模态图像配准已成为医学图像分析的一项重要研究内容。互信息是最为广泛使用的一种配准方法,其精度和鲁棒性都相对使人满意。通过实验证明了该方法的实用性,并且对其扩展形式如基于高阶熵的互信息和加入梯度的互信息等方法进行了分析,验证了空间信息在配准中的重要作用。
Muhimodality image registration is an important issue in medical image analysis. Mutual information is the most widely used method in multimodality registration. Both the accuracy and the robustness are satisfactory. The practicability of mutual information is validated with experiments. In addition, other extension similarity measures such as high-order mutual information and gradient information incorporated methods are analyzed, and the importance of spatial information is demonstrated.
出处
《计算机应用与软件》
CSCD
北大核心
2008年第3期145-146,170,共3页
Computer Applications and Software
关键词
互信息
图像配准
多模态
Mutual information Image registration Multimodality