独立分量分析及其在诱发电位提取中的应用  

Independent component analysis and its application in estimating VEP

作　　者：万相奎[1] 季忠[1] 梁小容[1] 秦树人[1] 

机构地区：[1]重庆大学机械学院测试中心,重庆400030

出　　处：《重庆大学学报（自然科学版）》2004年第8期7-10,共4页Journal of Chongqing University

基　　金：重庆市科学技术项目(7238)

摘　　要：独立分量分析能够在各信号频率混迭的情况下,从它们的线性混合信号中有效分离出原始信号,这为微弱电生理信号的提取提供了途径,受到生物信号处理研究人员的关注。作者研究分析了独立分量分析模型和一种基于负熵判据的定点快速ICA算法的实现,并将该算法运用于仿真诱发电位的提取中,从观测信号(混合信号)中有效地提取出视觉诱发电位。Original signals can be effectively estimated from their linear mixed signals by independent component analysis (ICA), even though their frequency components are overlapped, which provides a way to estimate the faint electrophysiological signals. And many researchers who study biomedical signal processing have paid great attention to the technique. We discuss the ICA model and the implement of a fixed-point fast ICA algorithm based on negentropy criterion. With the algorithm, we availably pick up the simulated visual evoked potential (VEP) from the mixing signals.

关 键 词：独立分量分析 视觉诱发电位 负熵 快速ICA算法 

分 类 号：TP38[自动化与计算机技术—计算机系统结构]