Scholarlink

In recent years, the arena of health imaging and image processing has witnessed remarkable advancements, particularly in the domain of lung disease detection. Medical imaging modalities such as X-rays, computed tomography (CT), and MRI are few indispensable tools for diagnosing various respiratory conditions. Image processing procedures play a vital role in enhancing the diagnostic capabilities of medical imaging. By applying algorithms and computational methods to analyze and interpret images, researchers and clinicians can extract valuable quantitative data, identify subtle abnormalities, and differentiate between healthy and diseased tissues. Many infected persons may be protected from lung disease if it is detected early. Chest X-rays and computed tomography (CT) pictures are the most widely utilized diagnostic techniques for lung disorders. Based on information from chest X-rays, this paper proposes a multi-classification DL model for identifying lung diseases. This article examines the performance of four architectures: VGGNet16+CNN, ResNet-152+CNN, EfficientNet-B7+CNN, and LungNet22. Through the use of publicly available digital chest X-ray datasets separated into three categories: Lung cancer, pneumonia, and normal, a comprehensive evaluation of numerous deep-learning architectures is provided. The experiment results show that the InceptionV3 plus CNN model outperforms the other three proposed models. The InceptionV3+CNN model obtained 97.55% accuracy, 97.43% sensitivity, 98.21% specificity, 98.07% F1-score, and 97.55% precision based on X-ray images.

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

Inception V3+CNN A Hybrid CNN Model for Lung Disease Detection Using Chest X-rays

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Anushree Raj,
  • M. Yogitha,
  • S. Deepthi

摘要

In recent years, the arena of health imaging and image processing has witnessed remarkable advancements, particularly in the domain of lung disease detection. Medical imaging modalities such as X-rays, computed tomography (CT), and MRI are few indispensable tools for diagnosing various respiratory conditions. Image processing procedures play a vital role in enhancing the diagnostic capabilities of medical imaging. By applying algorithms and computational methods to analyze and interpret images, researchers and clinicians can extract valuable quantitative data, identify subtle abnormalities, and differentiate between healthy and diseased tissues. Many infected persons may be protected from lung disease if it is detected early. Chest X-rays and computed tomography (CT) pictures are the most widely utilized diagnostic techniques for lung disorders. Based on information from chest X-rays, this paper proposes a multi-classification DL model for identifying lung diseases. This article examines the performance of four architectures: VGGNet16+CNN, ResNet-152+CNN, EfficientNet-B7+CNN, and LungNet22. Through the use of publicly available digital chest X-ray datasets separated into three categories: Lung cancer, pneumonia, and normal, a comprehensive evaluation of numerous deep-learning architectures is provided. The experiment results show that the InceptionV3 plus CNN model outperforms the other three proposed models. The InceptionV3+CNN model obtained 97.55% accuracy, 97.43% sensitivity, 98.21% specificity, 98.07% F1-score, and 97.55% precision based on X-ray images.