Xu Peng, Yao Dezhong, Luo Fen
School of Life Science and Technology, University of Electronic Science and Technology of China, Chengdu, China.
Sheng Wu Yi Xue Gong Cheng Xue Za Zhi. 2005 Aug;22(4):814-8.
The registration method based on mutual information is currently a popular technique for the medical image registration, but the computation for the mutual information is complex and the registration speed is slow. In engineering process, a subsampling technique is taken to accelerate the registration speed at the cost of registration accuracy. In this paper a new method based on statistics sample theory is developed, which has both a higher speed and a higher accuracy as compared with the normal subsampling method, and the simulation results confirm the validity of the new method.
基于互信息的配准方法是目前医学图像配准中的一种常用技术,但互信息的计算复杂且配准速度慢。在工程实践中,常采用子采样技术来加快配准速度,但以牺牲配准精度为代价。本文提出了一种基于统计抽样理论的新方法,与传统的子采样方法相比,该方法具有更高的速度和精度,仿真结果验证了新方法的有效性。