Noise-Resistant Algorithm for Sorting Cardio Complexes for the Task of Analyzing Statistical Characteristics of ECG

According to the WHO, cardiovascular diseases are the leading cause of death worldwide, so an accurate and timely diagnosis of cardiovascular diseases is an important task. One of the most common and effective methods for diagnosing CVD is to record an electrocardiogram using Holter monitoring. In order to significantly reduce the time required to decrypt the recording of ECS, cardiologists need special programs for automated analysis of the electrocardiogram. This is especially important for the long-term monitoring task. The program for the automated analysis of pacemakers should perform the clustering of cardiac complexes, thus dividing the electrocardiogram into groups of individual cardiac complexes. Only reference cardiocomplexes obtained by statistical averaging from each such group are subjected to further analysis. When registering an electrocardiogram, interference of various physical origins arises, artifacts that significantly complicate the analysis of ECS, therefore, automated analysis programs must also carry out preliminary processing of the electrocardiogram.