Scholarlink

The high frequency of kidney stones, as well as the accompanying disorders like pain, infection, and kidney damage, makes them one of the most pressing healthcare issues. Early detection is critical, but CT and ultrasound scan-based techniques are very resource intensive and even less predictable in a resource-constrained environment. This work develops and presents a federated reinforcement learning-assisted consumer system designed to enhance the detection efficiency to 94%, using CNN as feature extraction mechanisms, SVM to classify kidney stones, and privacy-preserving decentralized training mechanism through federated learning, to optimize the real-world feedback received by the reinforcement model. The FRL framework addresses data security, scalability, and improved diagnostic accuracy. Future research will include real-time optimization for resource-constrained environments, explainable AI, and multimodal imaging.

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

Deployment of Federated Learning Techniques for Privacy-Preserving and Distributed Kidney Stone Detection Using Medical Imaging Data

  • 章节

  • 发表日期：

    2026年

查看全文
  • V. Bharathi,
  • Madanapalli Hemanth Kumar,
  • Mohammed Shahrukh,
  • R. Prema,
  • Saket Ranjan Praveer,
  • Mamatha S. Upadhyariyadarshini

摘要

The high frequency of kidney stones, as well as the accompanying disorders like pain, infection, and kidney damage, makes them one of the most pressing healthcare issues. Early detection is critical, but CT and ultrasound scan-based techniques are very resource intensive and even less predictable in a resource-constrained environment. This work develops and presents a federated reinforcement learning-assisted consumer system designed to enhance the detection efficiency to 94%, using CNN as feature extraction mechanisms, SVM to classify kidney stones, and privacy-preserving decentralized training mechanism through federated learning, to optimize the real-world feedback received by the reinforcement model. The FRL framework addresses data security, scalability, and improved diagnostic accuracy. Future research will include real-time optimization for resource-constrained environments, explainable AI, and multimodal imaging.