基于胸导联ECG的心房颤动信号快速盲分离  

An efficient approach for blind separation of atrial fibrillation from precordial lead ECG

作　　者：丁志中[1] 叶中付[1] 汪一休[1] 

机构地区：[1]中国科学技术大学电子工程与信息科学系

出　　处：《中国科学技术大学学报》2007年第2期119-125,共7页JUSTC

基　　金：安徽省人才开发基金(2004Z025)资助

摘　　要：基于胸前6导联信号,提出采用非迭代的AMUSE算法实现房颤信号的快速盲分离,其特点是计算量和存储空间小,为实时分离系统和硬件系统的实现带来更多的有利条件.对真实心电信号的分离实验表明所提出的算法是有效的;且在弱房颤波6导联情况下,它比FastICA有更好的分离效果.An efficient and non-reiterative approach to separating AF blindly from precordial ECG signals was presented using AMUSE algorithm. The presented method has very low computation cost and simplifies electrode connection, which provides more favorable conditions for the implementation of real-time or hardware systems. Simulation based on real-world ECG signals shows that the approach is efficient, and it usually has better results of separation based on 6-lead ECG when atrial fibrillation signals are weak.

关 键 词：盲源分离 心房颤动 心电信号 多未知信号提取算法 

分 类 号：TN911.9[电子电信—通信与信息系统]