Scholarlink

<p>The integration of Internet of Things (IoT) in healthcare systems introduces critical security vulnerabilities, including data tampering, unauthorized access, and adversarial attacks, jeopardizing patient safety and regulatory compliance. This paper proposes&#xa0;BDCDC, a novel blockchain-deep learning framework designed to fortify data integrity and threat detection in IoT-imbued healthcare networks. BDCDC synergizes a blockchain layer, leveraging zero-knowledge protocols (ZKP) for privacy-preserving device authentication and smart contract-driven enhanced Proof-of-Work (ePoW) consensus, with a deep learning layer employing a Deep Variational Autoencoder (DVAE)&#xa0;for noise-resistant feature extraction and a Bidirectional Long Short-Term Memory (BiLSTM)&#xa0;model for real-time anomaly detection. To address scalability, an InterPlanetary File System (IPFS) offloads bulk data storage, minimizing blockchain overhead. Evaluated on the CICIDS-2017 and ToN-IoT datasets, BDCDC achieves exceptional performance, with 99.47–99.75% F1 scores for high-risk attacks (e.g., DDoS, MITM), near-zero false alarm rates (FAR &lt; 0.005%), and 99.06% accuracy, surpassing state-of-the-art methods. The framework reduces transaction latency by 2.1 × compared to traditional PoW and ensures GDPR/HIPAA compliance through decentralized, immutable audit trails. By harmonizing blockchain’s tamper-proof transparency with deep learning’s adaptive threat intelligence, BDCDC establishes a resilient, scalable defense against evolving cyber-physical threats in critical healthcare ecosystems. This work bridges the gap between theoretical security frameworks and practical deployment, offering a validated blueprint for safeguarding sensitive medical data in next-generation IoT infrastructures.</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

Blockchain and deep learning for improving impenetrable data passage in the IoT-imbued healthcare system

  • 期刊论文

  • 发表日期：

    2025年8月14日

查看全文
  • Isaac Osei Asante,
  • Libing Wu

摘要

The integration of Internet of Things (IoT) in healthcare systems introduces critical security vulnerabilities, including data tampering, unauthorized access, and adversarial attacks, jeopardizing patient safety and regulatory compliance. This paper proposes BDCDC, a novel blockchain-deep learning framework designed to fortify data integrity and threat detection in IoT-imbued healthcare networks. BDCDC synergizes a blockchain layer, leveraging zero-knowledge protocols (ZKP) for privacy-preserving device authentication and smart contract-driven enhanced Proof-of-Work (ePoW) consensus, with a deep learning layer employing a Deep Variational Autoencoder (DVAE) for noise-resistant feature extraction and a Bidirectional Long Short-Term Memory (BiLSTM) model for real-time anomaly detection. To address scalability, an InterPlanetary File System (IPFS) offloads bulk data storage, minimizing blockchain overhead. Evaluated on the CICIDS-2017 and ToN-IoT datasets, BDCDC achieves exceptional performance, with 99.47–99.75% F1 scores for high-risk attacks (e.g., DDoS, MITM), near-zero false alarm rates (FAR < 0.005%), and 99.06% accuracy, surpassing state-of-the-art methods. The framework reduces transaction latency by 2.1 × compared to traditional PoW and ensures GDPR/HIPAA compliance through decentralized, immutable audit trails. By harmonizing blockchain’s tamper-proof transparency with deep learning’s adaptive threat intelligence, BDCDC establishes a resilient, scalable defense against evolving cyber-physical threats in critical healthcare ecosystems. This work bridges the gap between theoretical security frameworks and practical deployment, offering a validated blueprint for safeguarding sensitive medical data in next-generation IoT infrastructures.