Heart Rate Monitoring System Using Feature Extraction in Electrocardiogram Signal by Convolutional Neural Network

Abstract

A new deep learning architecture, which is heart rate monitoring system using feature extraction in electrocardiogram signal by Convolutional Neural Network (CNN). Electrocardiogram based healthcare applications is presented in a federated context. The proposed system correctly diagnoses arrhythmias using an auto encoder and a classifier, both based on CNN. The module is provided to explain the classification findings in which the proposed classifier via employing an auto encoder and a classifier could check whether the rhythms of heart are normal, paced up or the heartbeat rate is irregular depending on the patient's situations. The module could offer the explanations of the classification findings in order to allow medical practitioners to quickly make the trustworthy judgments in preliminary diagnoses. Finally, the result shows that the proposed classifier could explain the classification for finding the two arrhythmias conditions which allow healthcare practitioners to rapidly make the correct conclusions.