Detecting cardiac arrhythmias with 1D convolutional neural network

Abstract

Arrhythmias, irregular heart rhythms, are a significant health concern, particularly among the elderly. The manual classification of electrocardiograms (ECGs), the standard method for detecting cardiac activity, is challenging due to the ECG signal's low amplitude, complexity, and non-linearity. Therefore, there is a pressing need for automated systems that can accurately identify abnormal heartbeats from large ECG datasets in healthcare. This study proposes a deep learning approach using Convolutional Neural Networks (CNNs) to automatically classify three types of ECG signals: normal (N), supraventricular ectopic beats (S), and ventricular ectopic beats (V). The proposed CNN model processes segments of ECG signals, each comprising 2000 samples, and their corresponding RR intervals from the MIT-BIH arrhythmia database using a multi-input structure. Through a five-fold cross-validation strategy, the model achieves an impressive accuracy of 99.28%. These results highlight the robust generalization performance of the proposed CNN model, indicating its potential as an effective tool to assist clinicians in diagnosing arrhythmias. The model's ability to accurately classify different types of arrhythmias from ECG signals demonstrates its potential for real-world clinical applications, paving the way for improved cardiac healthcare.