Scholarlink

The accurate and early diagnosis of Chronic Obstructive Pulmonary Disease (COPD) is pivotal for effective patient management and treatment. With the availability of big data for medical image analysis, there is an increasing emphasis on harnessing these advanced models for improved diagnostic capabilities. This study presents a methodology that leverages the power of fine-tuning pretrained ResNet50 and VGG19 neural networks for the efficient diagnosis of COPD from chest X-ray (CXR) images. Both models, originally designed for generic image classification tasks, were adapted to COPD diagnosis. Through systematic fine-tuning, the models were optimized to recognize subtle radiographic features indicative of the disease. Comparative evaluations highlighted the model's robustness, with ResNet50 showing slight superiority in accuracy and computational efficiency. The results underscore the potential of harnessing pretrained architectures, tailored through fine-tuning, as a promising avenue for the rapid and precise diagnosis of COPD using CXR images. This approach expedites the diagnostic process, and lays the groundwork for further advancements in AI-driven medical imaging for a myriad of lung diseases.

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

Optimizing ResNet50 and VGG19 Networks for Accurate COPD Diagnosis Through CXR Analysis

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Agughasi Victor Ikechukwu,
  • Sampoorna Bhimshetty

摘要

The accurate and early diagnosis of Chronic Obstructive Pulmonary Disease (COPD) is pivotal for effective patient management and treatment. With the availability of big data for medical image analysis, there is an increasing emphasis on harnessing these advanced models for improved diagnostic capabilities. This study presents a methodology that leverages the power of fine-tuning pretrained ResNet50 and VGG19 neural networks for the efficient diagnosis of COPD from chest X-ray (CXR) images. Both models, originally designed for generic image classification tasks, were adapted to COPD diagnosis. Through systematic fine-tuning, the models were optimized to recognize subtle radiographic features indicative of the disease. Comparative evaluations highlighted the model's robustness, with ResNet50 showing slight superiority in accuracy and computational efficiency. The results underscore the potential of harnessing pretrained architectures, tailored through fine-tuning, as a promising avenue for the rapid and precise diagnosis of COPD using CXR images. This approach expedites the diagnostic process, and lays the groundwork for further advancements in AI-driven medical imaging for a myriad of lung diseases.