Application of Modern Methods of Determining Characteristic Points of an Electrocardiogram for Ultra-High Resolution Electrocardiosignals

Coronary heart disease (CHD) is one of the most serious problems in the field of cardiology and medicine in general. In this regard, the method of ultra-high resolution electrocardiography (UHR ECG) is being actively developed, which is aimed at studying the fine structure of high-resolution electrocardiosignals (UHR ECS). The purpose of this method is to detect and register pathological changes in the cardiovascular system not only in areas that are traditionally analyzed in the processing of ECS, but also in areas that are considered artifacts. For effective automatic analysis of the amplitude and power characteristics of the ECG from the HF-channel, as well as determining the time when the cardiocycle of the signal from the LF-channel acquires a shape characteristic of the onset of CHD, it is important to be able to accurately determine the coordinates of the characteristic points of the ECG wave. The paper describes two methods for isolating PQRST-peaks: using a discrete wavelet transform (DWT) and a “naive” approach based on the idea of the standard cardiocycle shape, as well as an analysis of their effectiveness using an error matrix. Using the obtained data, a classifier based on the logistic regression method was built.