COVID-19 detection in X-ray images using convolutional neural networks

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Infectious Disease Diagnostics

Background:

• The COVID-19 pandemic necessitates rapid and accessible diagnostic tools.
• Current RT-PCR testing has limitations in terms of time and cost.
• Chest X-ray analysis shows promise for identifying COVID-19 indicators.

Purpose of the Study:

• To develop and evaluate a deep learning system for COVID-19 detection using Chest X-ray images.
• To assess the performance of VGG19 and U-Net models in classifying COVID-19 positive and negative cases.
• To explore the utility of lung segmentation and heatmap visualization in the diagnostic process.

Main Methods:

• Utilized deep learning models (VGG19, U-Net) for image classification.
• Implemented a preprocessing stage including lung segmentation to refine image data.
• Employed transfer learning for model training and heatmap visualization for result interpretation.

Main Results:

• The developed system achieved a COVID-19 detection accuracy of approximately 97%.
• Lung segmentation effectively removed irrelevant background information, potentially reducing bias.
• Heatmap visualization aided in understanding the model's classification decisions.

Conclusions:

• Deep learning models applied to Chest X-rays offer a highly accurate and efficient method for COVID-19 detection.
• This AI-driven approach presents a viable alternative to conventional diagnostic methods.
• Further development could enhance accessibility and speed of infectious disease diagnosis.