Fine Tuning of the Dynamic Low-pass Filter for Electromyographic Noise Suppression in Electrocardiograms

Abstract

In a series of publications, we have proposed and discussed the effectiveness of a dynamic low-pass filter for electromyographic (EMG) noise suppression in electrocardiograms (ECG). The goal of this study is to analyze the filter, and to suggest a better tuning for increasing the noise suppression and, at the same time, decreasing the signal distortion. The principle of the filter is the creation of a function called ‘wings’ for evaluation of the frequency spectra of the ECG waves. This function controls the dynamic cutoff frequency of an approximation procedure proposed by Savitzky and Golay, making it adjustable to the frequency spectra of ECG waves. The new way of forming the wings function permits: (i) stronger filtration of the low-frequency ECG components, (ii) reduced filtration in the transition zones of low to high frequency and vice versa (the highly diagnostic QRS onsets and offsets), and (iii) no filtration of the QRS zones of highest frequency. The newly suggested dynamic low-pass filtration of ECG performs better than the one suggested in the most recent publication.