Scholarlink

Recent breakthroughs in artificial intelligence (AI) play a crucial role in improving the precision, processing, visualization, and interpretation of chest X-ray (CXR) images for diagnosis in the field of medical imaging. CXR images are utilized for the medical evaluation of numerous disorders. This paper employs AI approaches to diagnose five distinct sorts of abnormalities like pneumonia, including Bacterial Pneumonia, Corona Virus Disease, Normal Pneumonia, Tuberculosis, and Viral Pneumonia. The images obtained are divided into segments to precisely identify the chest area and further improved employing gray-level modification techniques. The processed images are fed into pre-trained convolutional neural networks (CNNs), including Inception v3, DenseNet, MobileNet, Xception, VGG16, and an Ensemble model. In this stage, the transfer technique for learning is employed. In the ultimate fully linked layer, the automated characteristics are extracted. The accuracy of the Ensemble model was found to be 98.8%. The proposed methodology has the potential to serve as a significant advancement in the categorization of various pulmonary disorders that can be identified through the analysis of chest x-rays.

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

A Hybrid Deep Learning Approach to X-Ray Diagnosis of Lung Diseases

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • L. V. A. Priya Maddipati,
  • Durga Prasad Garapati,
  • Gutla Neerada Grace,
  • Kavya Sri Potthuri,
  • Bhanu Swetha Dadala,
  • K. Roosevelt

摘要

Recent breakthroughs in artificial intelligence (AI) play a crucial role in improving the precision, processing, visualization, and interpretation of chest X-ray (CXR) images for diagnosis in the field of medical imaging. CXR images are utilized for the medical evaluation of numerous disorders. This paper employs AI approaches to diagnose five distinct sorts of abnormalities like pneumonia, including Bacterial Pneumonia, Corona Virus Disease, Normal Pneumonia, Tuberculosis, and Viral Pneumonia. The images obtained are divided into segments to precisely identify the chest area and further improved employing gray-level modification techniques. The processed images are fed into pre-trained convolutional neural networks (CNNs), including Inception v3, DenseNet, MobileNet, Xception, VGG16, and an Ensemble model. In this stage, the transfer technique for learning is employed. In the ultimate fully linked layer, the automated characteristics are extracted. The accuracy of the Ensemble model was found to be 98.8%. The proposed methodology has the potential to serve as a significant advancement in the categorization of various pulmonary disorders that can be identified through the analysis of chest x-rays.