A Comparative Study of Deep Learning Models for ECG Signal-based User Classification

In recent years, Deep Learning (DL) has gained significant attention and has been extensively studied across various fields, including healthcare. In the context of healthcare, the timely detection of anomalies in Electrocardiogram (ECG) signals can play a crucial role in patient monitoring. Arrhythmia, characterized by irregular heart rate and abnormal heart rhythm, necessitates more than manual analysis of ECG signals for prompt identification of abnormalities. This paper presents a comprehensive review that focuses on the application of DL methods for the classification of ECG signals. Different DL techniques such as Convolutional Neural Network (CNN), INCEPTION, RESNET, Recurrent Neural Network (RNN), Long Short-Term Memory (LSTM), and Gated Recurrent Unit (GRU) have been considered in this study to detect five distinct types of arrhythmia. The results shows that the RESNET model surpasses all other existing DL models in terms of accuracy, achieving a remarkable 98.4% accuracy rate and identified that the ECG Signals can be used as a unique factor to identify user.