Parthajit Borah, Upasana Sarmah* and DK Bhattacharyya

Department of Computer Science and Engineering, Tezpur University, Napaam, Assam, India

*Corresponding Author: Upasana Sarmah, Department of Computer Science and Engineering, Tezpur University, Napaam, Assam, India.

Received: November 15, 2021; Published: March 17, 2022

Abstract

Recent history is not that generous and kind when it comes to viral infections from animal reservoirs to target humans. Re-emergence of mutating strains of such virus has only added more misery. In 2019, SARS-CoV2 had a daunting presence in and around the world, pausing grave threats to the perspective of global health, economy, livelihood and human race itself. Although numerous rigorous efforts are made to contain the contagious disease, there is a continuous steep rise in the number of clinically confirmed cases and fatalities. Medical facilities across the globe are in a crisis particularly when it comes to conducting adequate testing. RT-PCR (Reverse Transcription-Polymerase Chain Reaction) although reliable, consumes a lot of valuable time. Thus, an automated COVID-19 diagnosis strategy with efficient detection and minimum error is the need of the hour. Recently, Deep Learning techniques have become very popular for detecting COVID-19 using chest X-Ray images. In this paper, we propose a deep learning-based approach called DLCC to classify COVID-19 chest X-Rays with high accuracy. DLCC includes eight CNN architectures namely ResNet (18, 34, 50), AlexNet, VGG, and DenseNet (121, 161, 169) for best possible classification of diseased instances. All the models are fine-tuned using transfer learning. For the purpose of validation, a publicly available dataset containing four classes of chest radiograph, namely COVID-19, Lung opacity, Viral pneumonia and normal samples. From our study, it has been observed that DenseNet model gives the best performance in terms of accuracy (96.4%). This work might be used as a base to develop more effective CNN-based models for early detection of COVID-19.

Keywords: COVID-19; Pandemic; Virus; Genome; Strains; Coronavirus