Deep Convolutional Neural Networks for Accurate Diagnosis of COVID-19 Patients Using Chest X-Ray Image Databases from Italy, Canada, and the USA

Authors

Amgad A. Salama, Research and Development Center, Air Defense CollegeFollow
Samy H. Darwish, Electrical Engineering Department, Faculty of Engineering, Pharos UniversityFollow
Samir M. Abdel-Mageed, Physics Department, Faculty of Science, Alexandria UniversityFollow
Radwa A. Meshref, Medical Equipment Department, Faculty of Allied Medical Sciences, Pharos UniversityFollow
Ehab I. Mohamed, Alexandria UniversityFollow

Funder

The authors received no specific funding for this work.

Abstract

Introduction: Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), famously known as COVID-19, has quickly become a global pandemic. Chest X-ray (CXR) imaging has proven reliable, fast, and cost-effective for identifying COVID-19 infections, which proceeds to display atypical unilateral patchy infiltration in the lungs like typical pneumonia. We employed the deep convolutional neural network (DCNN) ResNet-34 to detect and classify CXR images from patients with COVID-19 and Viral Pneumonia and Normal Controls.

Methods: We created a single database containing 781 source CXR images from four different international sub-databases: the Società Italiana di Radiologia Medica e Interventistica (SIRM), the GitHub Database, the Radiology Society of North America (RSNA), and the Kaggle Chest X-ray Database for COVID-19 (n = 240), Viral Pneumonia (n = 274), and Normal Controls (n = 267). Images were resized, normalized, without any augmentation, and arranged in m batches of 16 images before supervised training, testing, and cross-validation of the DCNN classifier.

Results: The ResNet-34 had a diagnostic accuracy as of the receiver operating characteristic (ROC) curves of the true-positive rate versus the false-positive rate with the area under the curve (AUC) of 1.00, 0.99, and 0.99, for COVID-19 and Viral Pneumonia patient and Normal control CXR images; respectively. This accuracy implied identical high sensitivity and specificity values of 100, 99, and 99% for the three groups, respectively. ResNet-34 achieved a success rate of 100%, 99.6%, and 98.9% for classifying CXR images of the three groups, with an overall accuracy of 99.5% for the testing subset for diagnosis/prognosis.

Conclusions: Based on this high classification precision, we believe the output activation map of the final layer of the ResNet-34 is a powerful tool for the accurate diagnosis of COVID-19 infection from CXR images.

DOI

10.18297/jri/vol5/iss1/34

Creative Commons License

This work is licensed under a Creative Commons Attribution 4.0 License.

Recommended Citation

Salama, Amgad A.; Darwish, Samy H.; Abdel-Mageed, Samir M.; Meshref, Radwa A.; and Mohamed, Ehab I. (2021) "Deep Convolutional Neural Networks for Accurate Diagnosis of COVID-19 Patients Using Chest X-Ray Image Databases from Italy, Canada, and the USA," The University of Louisville Journal of Respiratory Infections: Vol. 5 : Iss. 1 , Article 34.
DOI: https://doi.org/10.18297/jri/vol5/iss1/34
Available at: https://ir.library.louisville.edu/jri/vol5/iss1/34