Deep Learning for Enhanced IoMT Security: A GNN-BiLSTM Intrusion Detection System
Yahya Rbah, Mohammed Mahfoudi, Mohammed Y. Fattah, Younes Balboul, Saïd Mazer, Moulhime Elbekkali, Benaissa Bernoussi · 2024
The growing number of Internet of Medical Things (IoMT) devices integrated into healthcare creates an expanding attack surface. Traditional intrusion detection methods struggle to adapt to this evolving landscape. Deep learning (DL) offers significant potential for improved threat detection in intrusion detection systems (IDS). This study presents a hybrid deep learning (DL) approach for IoMT security. The framework leverages a Graph Neural Network (GNN) and Bidirectional Long-Term Memory (Bi-LSTM) network for efficient and timely cyber threat detection within the IoMT system. We validated our method using the “IoT-Healthcare security” dataset. The proposed method achieved high accuracy (99.98%) and fast processing times (< 1.2 seconds) in threat detection, demonstrating its efficacy against prevalent threats.