Scholarlink

Deep learning (DL) has shown remarkable success in chest X-ray (CXR) disease diagnosis by enabling automated, high-accuracy interpretation of complex radio-graphic features. However, the centralized nature of data collection and model training introduces significant risks to patient privacy and data security, particu-larly in the context of regulatory frameworks such as HIPAA and GDPR. To ad-dress these concerns, blockchain technology offers decentralized, immutable, and transparent mechanisms for secure data exchange and model governance. This paper presents a comprehensive survey of blockchain-integrated DL frameworks designed for privacy-preserving CXR diagnosis. We analyze various architectures, including federated learning with blockchain, edge AI systems, and smart contract–driven diagnostic workflows. Key public CXR datasets and blockchain platforms are reviewed, alongside commonly used evaluation metrics. A detailed comparative analysis of 20 recent studies is provided, focusing on architectural design, privacy mechanisms, and performance outcomes. Additionally, we identify critical challenges such as scalability, real-time processing, model privacy leakage, and interoperability. The paper concludes with future research directions aimed at developing secure, scalable, and explainable AI systems for clinical deployment. This survey serves as a valuable reference for researchers and developers working at the intersection of medical imaging, deep learning, and blockchain technologies.

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

Advances in Blockchain-Enabled Deep Learning for Privacy-Preserving Chest X-Ray Diagnosis: Emerging Trends in Federated, Edge, and Explainable AI

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • R. Sriramkumar,
  • Joshuva Arockia Dhanraj,
  • Mitha Guru,
  • Mude Nagarjuna Naik,
  • M. Lakshmanan,
  • R. Rakshita

摘要

Deep learning (DL) has shown remarkable success in chest X-ray (CXR) disease diagnosis by enabling automated, high-accuracy interpretation of complex radio-graphic features. However, the centralized nature of data collection and model training introduces significant risks to patient privacy and data security, particu-larly in the context of regulatory frameworks such as HIPAA and GDPR. To ad-dress these concerns, blockchain technology offers decentralized, immutable, and transparent mechanisms for secure data exchange and model governance. This paper presents a comprehensive survey of blockchain-integrated DL frameworks designed for privacy-preserving CXR diagnosis. We analyze various architectures, including federated learning with blockchain, edge AI systems, and smart contract–driven diagnostic workflows. Key public CXR datasets and blockchain platforms are reviewed, alongside commonly used evaluation metrics. A detailed comparative analysis of 20 recent studies is provided, focusing on architectural design, privacy mechanisms, and performance outcomes. Additionally, we identify critical challenges such as scalability, real-time processing, model privacy leakage, and interoperability. The paper concludes with future research directions aimed at developing secure, scalable, and explainable AI systems for clinical deployment. This survey serves as a valuable reference for researchers and developers working at the intersection of medical imaging, deep learning, and blockchain technologies.