SARS-Net: COVID-19 detection from chest x-rays by combining graph convolutional network and convolutional neural...

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Infectious Diseases

Background:

• COVID-19 pandemic necessitates rapid diagnostic tools.
• Chest X-ray (CXR) imaging shows visual indicators of COVID-19.
• RT-PCR is accurate but resource-intensive for mass screening.

Purpose of the Study:

• To develop and evaluate SARS-Net, a novel deep learning model for COVID-19 detection.
• To leverage Convolutional Neural Networks and Graph Convolutional Networks for CXR analysis.
• To improve the accuracy and efficiency of COVID-19 diagnosis from medical images.

Main Methods:

• Developed a custom deep learning architecture named SARS-Net.
• Combined Graph Convolutional Networks (GCN) and Convolutional Neural Networks (CNN).
• Trained and validated the model on Chest X-ray images for COVID-19 diagnosis.

Main Results:

• SARS-Net achieved 97.60% accuracy on the validation set.
• The model demonstrated a sensitivity of 92.90% for COVID-19 detection.
• Quantitative analysis indicated superior performance compared to existing methods.

Conclusions:

• SARS-Net shows significant potential as a Computer-Aided Diagnosis (CADx) system for COVID-19.
• The model offers a promising approach for efficient and accurate screening of COVID-19 using CXR.
• Deep learning models can effectively assist in diagnosing infectious diseases from medical imaging.