A Deep Learning Based Attack Classification on IoMT Devices

Abstract

This study gives an in-depth evaluation of deep learning algorithms for threat classifying in the Internet of Medical Things, or IoMT, devices, a rapidly expanding subset of the Internet of Things (IoT) essential for modern healthcare systems. IoMT devices are more susceptible to cyberattacks as they gather, send, and process sensitive medical data, endangering patient security and privacy. Due to particular limitations like low processing power and strong privacy regulations, traditional security solutions usually need to catch up in IoMT circumstances. Consequently, deep learning algorithms offer viable substitutes for real-time, adaptive attack classification due to their capacity to identify intricate patterns in big datasets. Our evaluation and classification of state-of-the-art deep learning algorithms for IoMT security focuses on classification accuracy, computational efficiency, and threat adaption. We discuss significant challenges and opportunities, compare strategies, and assess success metrics. We examine CNNs, RNNs, and hybrid architectures, highlighting how well they can withstand various assaults. This survey will be a thorough resource to direct future studies and advancements in deep learning-based IoMT security.