Scholarlink

Detecting brain tumors early is crucial, for results; however, the utilization of Magnetic Resonance Imaging (MRI) in medical environments sparks worries about safeguarding patient data privacy and security concerns. This research proposed new methods for identifying brain tumors using federated learning. A technique allowing joint model training across organizations, without exposing sensitive patient information. The proposed model utilizes an upgraded neural network structure that integrates dropout and batch normalization methods to enhance the resilience and precision of the model. The experimentation includes creating MRI data to mimic patient situations and then training the model with datasets, from various origins while adhering to privacy rules. The assessment of collaborative learning model is done through metrics, like accuracy and precision as well as recall and F score. It’s worth mentioning that the proposed model reached a precision score of 0.89 when detecting tumors showing an ability to identify positive cases accurately while reducing false alarms. These findings illustrate that collaborative learning not only boosts the protection of medical information but also helps create effective and precise tools, for detecting brain tumors.

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

Toward Secure and Efficient Brain Tumor Detection: Federated Learning for Privacy-Preserving MRI Analysis

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Vikas Maurya,
  • Abdul Aleem

摘要

Detecting brain tumors early is crucial, for results; however, the utilization of Magnetic Resonance Imaging (MRI) in medical environments sparks worries about safeguarding patient data privacy and security concerns. This research proposed new methods for identifying brain tumors using federated learning. A technique allowing joint model training across organizations, without exposing sensitive patient information. The proposed model utilizes an upgraded neural network structure that integrates dropout and batch normalization methods to enhance the resilience and precision of the model. The experimentation includes creating MRI data to mimic patient situations and then training the model with datasets, from various origins while adhering to privacy rules. The assessment of collaborative learning model is done through metrics, like accuracy and precision as well as recall and F score. It’s worth mentioning that the proposed model reached a precision score of 0.89 when detecting tumors showing an ability to identify positive cases accurately while reducing false alarms. These findings illustrate that collaborative learning not only boosts the protection of medical information but also helps create effective and precise tools, for detecting brain tumors.