Efficient and privacy preserving two-party protocols for health monitoring system
Loading...
Date
2024-06
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
University of New Brunswick
Abstract
By examining electrocardiograms (ECG), respiration, and heart rates, doctors gain critical insights into a patient’s cardiovascular and respiratory health. Modern healthcare centers have adopted IoT-based wearable devices to collect this health data, enabling real-time feedback on patient health status.
As health data grows exponentially, healthcare centers turned to edge computing to solve the challenges of inadequate storage and processing power. However, this introduces privacy issues as user health information and health status are outsourced to edge servers. Hence, protecting the privacy of the health data, as well as the healthcare monitoring models, have become essential.
This report presents novel and effective protocols to perform training and inference while preserving privacy by using a two-server configuration, where private data is distributed by the data owners between two servers that do not collude by making use of secure two-party computation (2PC). We provide an analysis of the protocols' security, correctness, and performance.