An Arrhythmia Classification Approach via Deep Learning Using Single-lead ECG

Abstract

Abstract Arrhythmia, a frequently encountered and life-threatening cardiac disorder, can manifest as a transient or isolated event. Traditional automatic arrhythmia detection methods have predominantly relied on QRS-wave signal detection. Contemporary research has focused on the utilization of wearable devices for continuous monitoring of heart rates and rhythms through single-lead electrocardiogram (ECG), which holds the potential to promptly detect arrhythmias. However, in this investigation, we employed a convolutional neural network (CNN) to classify distinct arrhythmias without necessitating a signal detection step. The ECG data used in this study were sourced from publicly available databases. We randomly selected 5-second and 10-second segments of single-lead ECG data, accurately labeled for various arrhythmias, to train a one-dimensional CNN. In our experimental setup, the CNN model exhibited the capability to differentiate between Normal Sinus Rhythm (NSR) and various arrhythmias, including Atrial Fibrillation (AFIB), Atrial Flutter (AFL), Wolff-Parkinson-White syndrome (WPW), Ventricular Fibrillation (VF), Ventricular Tachycardia (VT), Ventricular Flutter (VFL), Mobitz II AV Block (MII), and Sinus Bradycardia (SB). Notably, both 10-second and 5-second ECG segments yielded a classification accuracy averaging 97.31%. This underscores the practicality of utilizing even brief 5-second recordings to detect arrhythmias in real-world scenarios.