A system for automatic processing, analysis and classification of ECG characteristics to determine various states of cardiac anomalies

Abstract

A systematic approach to the joint solution of problems of automatic processing, analysis, classification and measurement of characteristics of electrocardiogram signals is proposed, which ensures the elimination of one-sided solutions and organize a complex system consisting of various methods and algorithms. Algorithms have been developed for the proposed stages: preliminary processing, feature extraction, analysis of electrocardiograms, training, classification and recognition of characteristics of electrocardiograms, cardiac abnormalities, amplitude-time characteristics, diagnosis of diseases. A joint solution of discrete wavelet transform methods, artificial neural network, digital Butterworth filter, fast Fourier transform, Engelse and Zeelenberg segmentation, etc. All software products were implemented and tested using the MATLAB and PYTHON programming system.