A Cloud-Enabled Autonomous Telemedicine Platform for Continuous ECG Monitoring and Arrhythmia Classification

The Medical Internet of Things (MIoT), also known as telemedicine, has emerged as a promising paradigm for the continuous monitoring of patients' vital signs, while Artificial Intelligence (AI) has shown strong potential in healthcare applications, particularly for automated biosignal analysis. However, existing research has typically focused either on biosignal acquisition systems or on the development of deep learning models in isolation, with limited implementation in an integrated framework. This paper presents an autonomous IoT-based healthcare system for electrocardiogram (ECG) monitoring, integrated with an attention-based deep learning model for ECG signal classification. The developed cloud-integrated IoT platform enables the acquisition and storage of ECG signals while leveraging server-side computational resources for automated analysis and prediction. The proposed telemedicine system consists of a wearable device for ECG acquisition and a central server implementing deep learning models for five-class arrhythmia classification. The model achieved an accuracy of 99.0%, a precision of 93.6%, a recall of 90.8%, and an F1-score of 92.0%. To promote reproducibility and further research, the implementation of the proposed system is publicly available on GitHub (https://github.com/AlexTran1703/telemedicine-system). These findings demonstrate the feasibility of integrating IoT infrastructure with deep learning for remote ECG monitoring and automated arrhythmia classification, and support the potential of AI-enabled MIoT systems for scalable and efficient cardiovascular monitoring.