介绍了一种在标准 1 2导联心电图中消除噪声信号的正交化方法。使用一种基于奇异值分解 (SVD)的算法 ,将信号分解为两个时正交的子空间即一个包含了 ECG信号 ,另一个则包含了人工噪声 ,诸如基线漂移 (BW)和肌电信号 (EMG) ,这种方法利用了存在于 1 2导联心电图中的冗余。在完整丢失信道的重建方面该方法也适用。最后 ,给出了这种方法的在线实现

A SVD-Based Algorithm for ECG Noise Process Ai Chunting Li Chunguang Chen Zuobing Qian Xiangping (Department of Machine and Electronic Engineering, WUT,Wuhan 430070,China.)In this paper, we proposed an orthogonal method for noise signal elimination in standard 12-leads ECG. We used an algorithm based on single value decomposition to decomposed the signal into tow time- orthogonal subspaces: one contains ECG signal, the other contains artificial noise such as baseline wander and electromyography. This method used the redundancy existing in 12-lead ECG. This method also can used in reconstruction of complete lost channel. At last, we proposed on-line performance of this method.Single value decomposition; Electrocardiogram; Baseline wander; electromyography