Scholarlink

<p>The Internet of Medical Things (IoMT) has become a crucial part of modern healthcare systems, improving the quality of healthcare services by connecting medical devices to smart systems. However, this development comes with essential security risks, as the data managed is often sensitive and includes medical information that attackers can exploit, including Denial of Service (Dos)and Distributed Denial of Service (DDos) attacks. These Attacks can compromise data integrity, disrupt services, and hinder healthcare professionals from accessing critical patient information, posing severe risks to patient safety and clinical decision-making. This study aims to propose an Intrusion Detection System (IDS) model for IoMT systems using Deep learning (DL) techniques. It integrates Convolutional Neural Networks (CNN) with Gated Recurrent Unit (GRU) to improve attack detection accuracy. This model contributes to processing complex data by extracting spatial features using CNN and analyzing temporal patterns using gated repetition units (GRU). To address data imbalance, the modern SMOTE technique was used, which contributes to improving the model’s accuracy and reducing the possibility of bias in predictions. The model was tested on the modern medical dataset CICIoMT2024 to evaluate its ability to detect malicious activities with high accuracy, using metrics including precision, accuracy, recall, and F1 score. The results demonstrated a classification accuracy is 99.39%, The proposed IDS significantly enhances IoMT security by detecting and mitigating cyber threats effectively, ensuring the protection of sensitive medical information.</p>

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

Intrusion detection in IoMT systems using hybrid CNN-GRU deep learning model

  • 期刊论文

  • 发表日期：

    2026年6月17日

查看全文
  • Shoog Alshehri,
  • Asma Alnemari

摘要

The Internet of Medical Things (IoMT) has become a crucial part of modern healthcare systems, improving the quality of healthcare services by connecting medical devices to smart systems. However, this development comes with essential security risks, as the data managed is often sensitive and includes medical information that attackers can exploit, including Denial of Service (Dos)and Distributed Denial of Service (DDos) attacks. These Attacks can compromise data integrity, disrupt services, and hinder healthcare professionals from accessing critical patient information, posing severe risks to patient safety and clinical decision-making. This study aims to propose an Intrusion Detection System (IDS) model for IoMT systems using Deep learning (DL) techniques. It integrates Convolutional Neural Networks (CNN) with Gated Recurrent Unit (GRU) to improve attack detection accuracy. This model contributes to processing complex data by extracting spatial features using CNN and analyzing temporal patterns using gated repetition units (GRU). To address data imbalance, the modern SMOTE technique was used, which contributes to improving the model’s accuracy and reducing the possibility of bias in predictions. The model was tested on the modern medical dataset CICIoMT2024 to evaluate its ability to detect malicious activities with high accuracy, using metrics including precision, accuracy, recall, and F1 score. The results demonstrated a classification accuracy is 99.39%, The proposed IDS significantly enhances IoMT security by detecting and mitigating cyber threats effectively, ensuring the protection of sensitive medical information.