PREPROCESSING AND ANALYSIS OF ELECTROCARDIOGRAM SIGNALS

Abstract

Modern functional diagnostics of cardiovascular diseases have large number of different instrumental research methods, but one of the most common is electrocardiography (ECG). ECG signals are easy to measure, store electronically, and process using digital methods. These and some other features of the ECG create certain prospects for their application in diagnostic systems.

ECG signals are significantly dependent on the actual psychological and physical condition of the person. This greatly complicates the signal processing algorithms.  ECG signals are also affected by external noise sources: electrode contact noise; short-term noise resulting from patient movement; interference created by the electromagnetic field from surrounding electrical devices, etc. Therefore, the original signals must undergo preliminary pre-processing, which includes filtering.

Filtering ECG signals allows you to remove extraneous noise while preserving key signal features, such as R peaks, RR intervals, and heart rate. Removing noise from an ECG signal improves its accuracy and allows using this data for heart rate variability (HRV) analysis and signal frequency analysis. Butterworth, Chebyshev, and elliptic filters were used in the work.

Frequency analysis of ECG signals using fast Fourier transform (FFT) and wavelet analysis allowed us to study the frequency composition of signals in detail and identify their features. Visualization of the results of filtering and analysis helped to clearly assess the effectiveness of filtering and interpret the obtained data.

The studies considered show the high efficiency of using the developed software for signal analysis and open the way for further research aimed at extracting additional information from ECG signal, which would serve as the basis for early diagnosis of cardiovascular diseases.