RADAI: A Deep Learning-Based Classification of Lung Abnormalities in Chest X-Rays.

Area of Science:

• Medical Imaging
• Artificial Intelligence in Healthcare
• Radiology

Background:

• Chest X-rays are increasingly vital for diagnosing lung conditions, as noted by the WHO.
• Interpreting chest X-rays is challenging, often leading to diagnostic delays and errors.
• Automated analysis of medical images, including chest X-rays, shows significant promise.

Purpose of the Study:

• To develop RadAI, an artificial intelligence model for detecting lung abnormalities in chest X-rays.
• To enable RadAI to generate detailed reports for identified abnormalities.
• To enhance the accuracy and efficiency of chest X-ray interpretation.

Main Methods:

• Fine-tuning three deep learning models: Feature-selective and Spatial Receptive Fields Network (FSRFNet50), ResNext50, and ResNet50.
• Utilizing convolutional neural networks (CNNs) for automated medical image analysis.
• Comparing model performance using metrics such as accuracy, precision, recall, and F1-score.

Main Results:

• The developed RadAI model demonstrated high performance in detecting lung abnormalities.
• RadAI accurately identifies four distinct types of lung abnormalities from chest X-rays.
• The model's performance indicates its potential to aid radiologists in accurate interpretation.

Conclusions:

• RadAI significantly enhances the accuracy and efficiency of chest X-ray interpretation.
• The tool supports timely and reliable diagnosis of lung abnormalities.
• RadAI serves as a valuable assistant for radiologists in clinical practice.