Scholarlink

Artificial intelligence (AI) is increasingly becoming a part of everyday lives, driven by the large amounts of data available for application in various fields. Accordingly, new possibilities have appeared in the field of medical science. Deep learning models could help reduce manual work and, in some cases, make more accurate diagnostic decisions. The global COVID-19 pandemic has posed a huge challenge to healthcare systems in many countries, putting healthcare personnel under enormous pressure in previously unseen working conditions. In such circumstances, the existence of AI classification models could benefit both medical staff and patients. This paper compares different convolutional neural networks architectures for pneumonia detection, using a dataset which originally consisted of 5856 X-ray images. To prevent overfitting, data augmentation was performed, as well as weight regularization technique known as L2 regularization. Different optimization algorithms such as Adam and mini-batch were also compared.

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

Detection of Pneumonia Using VGG-16, MobileNetV2, and ResNet50 Convolutional Neural Networks

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Dženana Pačariz Ramaš,
  • Tatjana Konjić,
  • Azur Ramaš,
  • Amira Šerifović-Trbalić

摘要

Artificial intelligence (AI) is increasingly becoming a part of everyday lives, driven by the large amounts of data available for application in various fields. Accordingly, new possibilities have appeared in the field of medical science. Deep learning models could help reduce manual work and, in some cases, make more accurate diagnostic decisions. The global COVID-19 pandemic has posed a huge challenge to healthcare systems in many countries, putting healthcare personnel under enormous pressure in previously unseen working conditions. In such circumstances, the existence of AI classification models could benefit both medical staff and patients. This paper compares different convolutional neural networks architectures for pneumonia detection, using a dataset which originally consisted of 5856 X-ray images. To prevent overfitting, data augmentation was performed, as well as weight regularization technique known as L2 regularization. Different optimization algorithms such as Adam and mini-batch were also compared.