Framework for Processing Medical Data and Secure Machine Learning in Internet of Medical Things

Abstract

The term «Internet of Medical Things» (IoMT) refers to a set of devices and technologies for remote monitoring of patients’ health using wearable devices. One primary problem with patient’s data is ensuring privacy when it is transmitted over open communication channels and stored in cloud systems. A whole range of different approaches to these issues are available. However, when it comes to millions of IoT devices, technologies that have already become classic for Internet resources are not suitable in many aspects at once. The aim of this work is to develop methods and protocols for secure interaction between portable diagnostic devices and cloud services for the analysis and processing of medical data in the Internet of Medical Things networks. The work considered existing technologies and solutions for ensuring security in IoMT networks and personalized medicine systems; also, it focused on secure machine learning methods. Previous studies have emphasized attribute-based encryption (ABE) as a prospective method for data privacy and security. These algorithms solve many problems for IoMT applications: patient’s data confidentiality, flexible key management, fine-grained access control mechanisms, and user control over data. We have proposed a framework for processing patient data from portable diagnostic devices using ABE methods.