An improved denoise method based on independent component analysis to remove ECG noise in surface EMG signal

Abstract

An improved denoise method based on independent component analysis is discussed in this article to remove ECG noise in surface EMG signal. In order to prevent loss of useful EMG signal, low pass filter is used to process the ECG components decomposed from the original surface EMG signal. And then, project the processed ECG components back to the original space, subtract them from the original signal. Experiment on true 16-channel mono-polar surface EMG signal indicates this method can achieve satisfying results.

本文提出了一種基于獨立成份分析(ICA)的改進方法,去除肌電信號中的心電噪聲。該方法基于獨立成份分析算法,在肌電成份中分解出心電獨立成份:對分解出的心電獨立成份進行低通濾波處理,進一步降低了其中肌電成份;而將其反向投影回原始信號空間,并從原始信號中除去,以便在最大程度上消除肌電信號中的心電干擾成份。真實16通道單極性肌電信號的處理和分析表明,該方法取得了較好的效果。