基于SPM的功能磁共振成像图像头动校正配准方法的分析与改进

Analysis and Improvement of Functional MRI Image Registration Method Based on SPM

功能磁共振成像作为一种新兴的神经影像学成像技术,其重要性受到神经、认知和心理学等领域的广泛关注。本文针对功能磁共振成像数据预处理过程中的头动校正环节,较为全面地阐述了国际知名脑功能成像软件包SPM的配准原理,并在此基础上提出了一种改进的配准方法,通过由粗到细的处理有效地提高了配准的准确度。此外,将差值平方和与归一化互信息的比值作为配准测度的思想弥补了单用一种测度的局限性,增加了配准过程的鲁棒性和可靠性。研究结果显示该方法具有一定的实际意义和应用价值。

As a new neuroimaging method, functional magnetic resonance（f MRI） with its importance of data processing has been widely recognized by neurology and cognitive psychology. Focusing on the realignment section in f MRI image preprocessing, this paper comprehensively describes the registration principle of SPM, an internationally-known software package which is specially designed for cerebral function imaging. An improved registration method is presented which effectively increases the accuracy. In addition, choosing the ratio of SSD and NMI as the registration measure can compensate for the limitation of using single measurement, which improves the robustness and reliability of the registration process. Experimental results prove the feasibility of this method.