Scholarlink

<p>The increasing prevalence of cyber threats in healthcare necessitates robust security measures to protect sensitive medical data. This research presents a hybrid security framework integrating blockchain for decentralized, immutable data storage and an intrusion detection system (IDS) leveraging a multi-attention deep convolutional recurrent neural network (MA-DeepCRNN) model for advanced threat detection. The proposed IDS combines convolutional neural networks (CNNs) for spatial feature extraction, recurrent neural networks (RNNs) for temporal pattern recognition, and an attention mechanism to enhance critical data representation. The model is evaluated using the CICIoMT 2024 benchmark dataset. The blockchain architecture achieves a block creation time of 10&#xa0;s, improving significantly over Bitcoin (~ 600&#xa0;s) and Ethereum (~ 15&#xa0;s), while increasing throughput to  ~ 182 transactions per second. Security analysis indicates a low transaction reversal probability of &lt; 0.1%. The IDS demonstrates high classification performance, achieving 99.49% accuracy in binary classification, 99.12% in multiclass (6-class) classification, and 98.56% in large-scale (19-class) classification. Comparative analysis with state-of-the-art approaches highlights improvements in accuracy and F1-score by 3.44 and 3.71%, respectively, for intrusion detection in Internet of Medical Things (IoMT) systems. These results underscore the effectiveness of the proposed framework in enhancing security, scalability, and real-time threat detection in healthcare environments.</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

Multi-attention DeepCRNN: an efficient and explainable intrusion detection framework for Internet of Medical Things environments

  • 期刊论文

  • 发表日期：

    2025年4月5日

查看全文
  • Nikhil Sharma,
  • Prashant Giridhar Shambharkar

摘要

The increasing prevalence of cyber threats in healthcare necessitates robust security measures to protect sensitive medical data. This research presents a hybrid security framework integrating blockchain for decentralized, immutable data storage and an intrusion detection system (IDS) leveraging a multi-attention deep convolutional recurrent neural network (MA-DeepCRNN) model for advanced threat detection. The proposed IDS combines convolutional neural networks (CNNs) for spatial feature extraction, recurrent neural networks (RNNs) for temporal pattern recognition, and an attention mechanism to enhance critical data representation. The model is evaluated using the CICIoMT 2024 benchmark dataset. The blockchain architecture achieves a block creation time of 10 s, improving significantly over Bitcoin (~ 600 s) and Ethereum (~ 15 s), while increasing throughput to  ~ 182 transactions per second. Security analysis indicates a low transaction reversal probability of < 0.1%. The IDS demonstrates high classification performance, achieving 99.49% accuracy in binary classification, 99.12% in multiclass (6-class) classification, and 98.56% in large-scale (19-class) classification. Comparative analysis with state-of-the-art approaches highlights improvements in accuracy and F1-score by 3.44 and 3.71%, respectively, for intrusion detection in Internet of Medical Things (IoMT) systems. These results underscore the effectiveness of the proposed framework in enhancing security, scalability, and real-time threat detection in healthcare environments.