多个空间高斯源信号情况下成组三维图像特征提取方法

Group Three-Dimensional Image Feature Extraction Methodof Multiple Spatial Gaussian Source Signals

摘要:

针对功能磁共振成像数据中含有多个高斯信号源的盲源分离问题,介绍了一种成组典型相关分析方法(Group BSS-CCA)。这个方法的组分析框架参照了GIFT工具箱中的Group ICA算法,具体的典型相关分析方法应用的是Friman等人提出的BSS-CCA算法。以验证该方法的有效性为目的,设计了仿真实验；结果表明,该方法能较好识别出混合在人脑功能磁共振成像数据的2个空间高斯信号。Group BSS-CCA算法对研究人脑的功能磁共振成像数据具有较高的实用价值。

Abstract:

A new method was introduced for the problem of blind source separation with multiple Gaussian signal sources in functional magnetic resonance imaging data. The group inference framework of this method refered to the group ICA al? gorithm in the GIFT toolbox. The specific canonical correlation analysis method was the BSS-CCA algorithm proposed by Friman et al. The simulation results showed that the method could well identify two spatial Gaussian signals mixed with hu? man brain magnetic resonance imaging data. The results showed that the method could be used to verify the effectiveness of the method. Group BSS-CCA had a high practical value in the study of functional magnetic resonance imaging data of human brain.