Scholarlink

<p>The emergence of Healthcare 5.0 signifies a major shift in the medical field, driven by the integration of advanced technologies to transform patient care. This paper introduces a novel framework that combines Blockchain technology with Federated Learning (FL) to tackle key challenges in data security, privacy, and scalability within healthcare systems. The proposed approach focuses on classifying and detecting various lung diseases using FL. Local training algorithms, asynchronous FL, model aggregation, and enhanced FL were applied to datasets from four hospitals. The trained global model was then used to process over 12,500 chest images from the Kaggle repository, employing a stratified sampling technique with a 70:15:15 ratio for training, validation, and testing. Performance metrics including accuracy, precision, recall, and F1-Score were 93%, 0.93, 93.2%, and 0.93, respectively. The comprehensive simulations and comparative analyses show that this model significantly surpasses traditional methods in both security and predictive accuracy. This study highlights the potential of advanced, human-centric technologies in Healthcare 5.0, paving the way for more personalized, secure, and intelligent healthcare services.</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

BCT-FLHD: A blockchain-enabled federated learning framework for healthcare 5.0 disease detection

  • 期刊论文

  • 发表日期：

    2025年6月14日

查看全文
  • Kajal Tiwari,
  • Sanjay Kumar

摘要

The emergence of Healthcare 5.0 signifies a major shift in the medical field, driven by the integration of advanced technologies to transform patient care. This paper introduces a novel framework that combines Blockchain technology with Federated Learning (FL) to tackle key challenges in data security, privacy, and scalability within healthcare systems. The proposed approach focuses on classifying and detecting various lung diseases using FL. Local training algorithms, asynchronous FL, model aggregation, and enhanced FL were applied to datasets from four hospitals. The trained global model was then used to process over 12,500 chest images from the Kaggle repository, employing a stratified sampling technique with a 70:15:15 ratio for training, validation, and testing. Performance metrics including accuracy, precision, recall, and F1-Score were 93%, 0.93, 93.2%, and 0.93, respectively. The comprehensive simulations and comparative analyses show that this model significantly surpasses traditional methods in both security and predictive accuracy. This study highlights the potential of advanced, human-centric technologies in Healthcare 5.0, paving the way for more personalized, secure, and intelligent healthcare services.