Scholarlink

Intrusion detection within Internet of Medical Things (IoMT) environments is complicated by the diversity of communication protocols and the continuous emergence of sophisticated security threats. This research presents a hybrid deep learning framework that harnesses the complementary capabilities of Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) networks to enhance intrusion detection in healthcare IoMT environments. Our CNN-LSTM architecture employs CNN for extracting spatial features and LSTM for modeling temporal dependencies, specifically designed for IoMT data. Evaluated using the CICIoMT2024 dataset, covering Bluetooth, WiFi, and MQTT protocols with 18 attack types grouped into five classes, the model achieved an accuracy of 86.24% in multi-class classification, along with strong precision (0.865), recall (0.863), and F1-score (0.863) metrics.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Towards Reliable and Secure IoMT: A Deep Learning Perspective on Cyber-Physical Threats

  • 章节

  • 发表日期：

    2026年

查看全文
  • Hafida Assmi,
  • Said Jabbour,
  • Azidine Guezzaz

摘要

Intrusion detection within Internet of Medical Things (IoMT) environments is complicated by the diversity of communication protocols and the continuous emergence of sophisticated security threats. This research presents a hybrid deep learning framework that harnesses the complementary capabilities of Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) networks to enhance intrusion detection in healthcare IoMT environments. Our CNN-LSTM architecture employs CNN for extracting spatial features and LSTM for modeling temporal dependencies, specifically designed for IoMT data. Evaluated using the CICIoMT2024 dataset, covering Bluetooth, WiFi, and MQTT protocols with 18 attack types grouped into five classes, the model achieved an accuracy of 86.24% in multi-class classification, along with strong precision (0.865), recall (0.863), and F1-score (0.863) metrics.