Algorithm for EMG noise level approximation in ECG signals

摘要

In this paper, we introduce an approach for Electromyogram (EMG) noise level approximation in Electrocardiogram (ECG) signals. The stationary wavelet transform (SWT) is used to find efficient translation-invariant approximation of EMG noise. This is accomplished in the form of reference signal extracted as an estimation of the signal quality vs. EMG noise. The reference signal is built and then normalized after considering different heart rates and rhythms which increases its robustness and reliability to give accurate results regardless of input signal rhythm. Additionally, four applications of the extracted reference signal are suggested in this paper.