Abstract

Continuous ECG monitoring is crucial for early heart disease predictions; however, traditional data management and model training may risk the exposure of sensitive patient information. We propose a secure ECG monitoring system by integrating federated learning with blockchain technology. A mobile app helps patients monitor their heart condition and the doctors provide feedback, all connected through a private blockchain server. In addition, Federated learning allows secure and scalable ECG classification model training across a large number of patient devices. In our simulated experiment, the federated learning system managed up to 500 clients, which demonstrated its capability to handle large amounts of clients.

Share

COinS