Deep Learning-based Decision Support System for classification of COVID-19 and Pneumonia patients

Misha Urooj Khan (Department of Electronics Engineering University of Engineering and Technology Taxila Pakistan)
Zubair Saeed (Department of Computer Engineering University of Engineering and Technology Taxila Pakistan)
Ali Raza (Department of Computer Engineering University of Engineering and Technology (UET) Taxila, Pakistan)
Zeeshan Abbasi
Syeda Zuriat-e-Zehra Ali (Department of Electrical Engineering University of Engineering and Technology Taxila Pakistan)
Hareem Khan (Department of Computer Engineering University of Engineering and Technology (UET) Taxila, Pakistan)

JAREE (Journal on Advanced Research in Electrical Engineering)
Vol 6, No 1 (2022): April

The fast spread of Coronavirus (COVID-19) poses a huge risk to people all around the world. Recently, COVID-19 testing kits have been unavailable due to rise in effected people and large demand of tests. Keeping the urgency of the situation in mind, an automatic diagnosis method for early detection of COVID-19 is needed. The proposed deep learning decision support system (DSS) for COVID-19 employs MobileNet v2 Deep learning (DL) model for effective and accurate detection. Here we collected Cough auscultations through self-designed digital sensor. The primary experimental results show that the maximum accuracy for training is around 99.91%, and the maximum accuracy for validation is 98.61%, with 97.5% precision, 98.5%recall, and 98% F1-score. The Deep Learning-based model described here strives for similar performance to medical professionals and can help pulmonologist/radiologists increase their working productivity.