Scholarlink

Pneumonia is a substantial risk for global health. The prompt and precise diagnosis is important and very crucial. This study centers on the diagnosis of pneumonia through the analysis of chest X-ray images using sophisticated deep learning models within a federated learning framework. We utilized three convolutional neural networks: VGG-16, ResNet-50, and InceptionV3 with a federated learning approach. The models underwent training and evaluation, resulting in accuracies of 90%, 87%, and 89% respectively. Implementing federated learning in this study is essential for improving data privacy and security. Conventional centralized machine learning methods typically involve gathering data in a single storage location, which can lead to worries about patient privacy and data security breaches. Federated learning allows for the training of models on multiple decentralized devices or servers that hold local data samples, without the need to exchange the data. This paradigm not only ensures the preservation of data privacy by storing the data on local devices, but it also complies with regulatory requirements and minimizes the likelihood of data exposure.

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

Diagnosis of Pneumonia from Chest X-Ray Images Using Federated Learning

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Asadi Srinivasulu,
  • Saurabh Kumar,
  • Rohit Chowdhury,
  • Rahul Mahto,
  • Anupam Agrawal

摘要

Pneumonia is a substantial risk for global health. The prompt and precise diagnosis is important and very crucial. This study centers on the diagnosis of pneumonia through the analysis of chest X-ray images using sophisticated deep learning models within a federated learning framework. We utilized three convolutional neural networks: VGG-16, ResNet-50, and InceptionV3 with a federated learning approach. The models underwent training and evaluation, resulting in accuracies of 90%, 87%, and 89% respectively. Implementing federated learning in this study is essential for improving data privacy and security. Conventional centralized machine learning methods typically involve gathering data in a single storage location, which can lead to worries about patient privacy and data security breaches. Federated learning allows for the training of models on multiple decentralized devices or servers that hold local data samples, without the need to exchange the data. This paradigm not only ensures the preservation of data privacy by storing the data on local devices, but it also complies with regulatory requirements and minimizes the likelihood of data exposure.