On the Implementation of a Post-Pandemic Deep Learning Algorithm Based on a Hybrid CT-Scan/X-ray Images...

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Pulmonology

Background:

• Accurate and rapid diagnosis of lung diseases is crucial.
• Interpreting medical lung images can be challenging for physicians, leading to diagnostic errors.
• Artificial intelligence, particularly deep learning, offers potential solutions for improving diagnostic accuracy.

Purpose of the Study:

• To develop and evaluate a deep learning architecture for classifying lung diseases from medical images.
• To compare the proposed model's performance against existing pneumonia detection techniques.
• To create an accurate computer-aided system for analyzing radiographic and CT lung images.

Main Methods:

• A deep learning architecture based on EfficientNetB7 was constructed.
• The model was trained to classify medical X-ray and CT images into three categories: common pneumonia, coronavirus pneumonia, and normal cases.
• Performance was evaluated by comparing accuracy with recent pneumonia detection methods.

Main Results:

• The proposed EfficientNetB7 model achieved high predictive accuracy for pneumonia detection.
• Accuracy rates were 99.81% for radiography and 99.88% for CT scans.
• The system demonstrated robust and consistent feature extraction for pneumonia identification.

Conclusions:

• The developed deep learning system provides an accurate computer-aided tool for lung disease analysis.
• The promising results indicate potential for improved diagnosis and decision-making in lung disease management.
• This approach can aid in the timely identification of emerging lung diseases.