中国组织工程研究

中国组织工程研究 ›› 2010, Vol. 14 ›› Issue (52): 9772-9776.doi: 10.3969/j.issn.1673-8225.2010. 52.021

• 骨与关节图像与影像 bone and joint imaging • 上一篇    下一篇

SPM软件包头动校正原理的分析与评价

王  超1，杜剑平2，闫  镔1，王大会2，史大鹏2   

  1. 1 信息工程大学信息工程学院，河南省郑州市 450002；2 河南省人民医院放射科，河南省郑州市 450002
  • 出版日期:2010-12-24 发布日期:2010-12-24
  • 通讯作者: 闫镔，博士，副教授，信息工程大学信息工程学院，河南省郑州市 450002 ybspace2000@sina.com
  • 作者简介:王超，男，1985年生，湖南省常德市人，汉族，2008年信息工程大学毕业，主要从事检测技术与自动化方面的研究。 charleswang0100@gmail.com
  • 基金资助:

    本课题受河南省科技发展计划项目(072300450240)资助。　

Analysis and evaluation on the principle of realignment of the statistical parametric mapping software package

Wang Chao1, Du Jian-ping2, Yan Bin1, Wang Da-hui2, Shi Da-peng2   

  1. 1 Information Engineering College, Information Engineering University, Zhengzhou  450002, Henan Province, China; 2 Department of Radiology, Henan People's Hospital, Zhengzhou  450002, Henan Province, China
  • Online:2010-12-24 Published:2010-12-24
  • Contact: Yan Bin, Doctor, Associate professor, Information Engineering College, Information Engineering University, Zhengzhou 450002, Henan Province, China ybspace2000@sina.com
  • About author:Wang Chao, Information Engineering College, Information Engineering University, Zhengzhou 450002, Henan Province, China charleswang0100@gmail.com
  • Supported by:

    the Science and Technology Development Program of Henan Province, No. 072300450240*

PDF

891

被引次数

7

摘要：

背景：头动校正是功能磁共振图像处理的一个重要步骤，也是目前应用广泛的SPM软件的第一步，其效果直接关系到结果的可靠性。
目的：分析SPM软件包头动校正处理的基本原理和关键算法，包括刚体变换、高斯牛顿迭代算法和B样条插值算法。
方法：使用SPM2软件包对一组实际数据进行了处理，并结合头动校正的原理，提出了头动校正效果的评价方法：残差图与残差平方和。
结果与结论：使用高斯牛顿迭代算法得到优化的刚体变换参数，用B样条插值算法对须校正的图像进行重采样得到校正后图像，实现图像序列的头动校正。实验结果表明，评价指标能较好衡量算法效果。

关键词: 头动校正, SPM, 刚体变换, 高斯牛顿迭代算法, B样条插值, 残差图

Abstract:

BACKGROUND: Realignment is an important process in functional magnetic resonance imaging (fMRI) and the first step in the image preprocess of the widely used software package statistical parametric mapping (SPM). The reliability of the results is determined by the effect of realignment process.
OBJECTIVE: To analyze the basic principle and the key arithmetic in the process of realignment of SPM, including rigid body transformation, Gauss-Newton optimization and B-spline Interpolation.
METHODS: A set of real data were processed utilizing SPM2. Combined with the principle of the realignment, a way to evaluate the effect of the realignment is proposed: measure map and measure quality.
RESULTS AND CONCLUSION: Gauss-Newton optimization obtained optimized rigid body transformation parameters; B-spline Interpolation was used to obtain corrected images. Results show that these evaluation indexes can preferably evaluate the principle effect.

中图分类号: