Characterizations of critical heart disease in ECG signal features through recurrence plots as for medical imaging diagnostics

Abstract

this paper presents the realization of recurrence plots as a tool for characterizations of critical heart diseases in electrocardiogram (ECG) signal features. The technique is typically a conversion of a time-domain ECG signal to a colored image in a particular pattern of recurrence plots in order to indicate normal or abnormal heart conditions. The proposed technique is distinctive to other previously suggested approaches since the investigation directly focuses on the recurrence plot of each lead that significantly reveals the cause and type of heart diseases. Fifteen-lead ECG signals provided by MIT-BIH databases were utilized, including of one normal and particularly selected eight critical anomalies, involving Bundle Branch Block, Cardiomyopathy, Dysrhythmia, Myocardial Hypertrophy, Myocardial Infarction, Myocarditis, Valvular Heart Disease, and Unstable Angina. Simulations were performed in MATLAB, and the results reveal apparent difference in recurrence plots even though there is little change in time domain. This paper therefore proposes an alternative tool for characterizing critical heart diseases in ECG signal features for further utilization of medical imaging diagnostics.