ELUCNN for explainable COVID-19 diagnosis

Shui-Hua Wang; Suresh Chandra Satapathy; Man-Xia Xie; Yu-Dong Zhang

doi:10.1007/s00500-023-07813-w

ELUCNN for explainable COVID-19 diagnosis

Soft comput. 2023 Jan 13:1-17. doi: 10.1007/s00500-023-07813-w. Online ahead of print.

Authors

Shui-Hua Wang^#^{1

2

3}, Suresh Chandra Satapathy^#⁴, Man-Xia Xie⁵, Yu-Dong Zhang^{1

2

3}

Affiliations

¹ School of Computer Science and Technology, Henan Polytechnic University, Jiaozuo, 454000 Henan People's Republic of China.
² School of Computing and Mathematical Sciences, University of Leicester, Leicester, LE1 7RH UK.
³ Department of Information Systems, Faculty of Computing and Information Technology, King Abdulaziz University, Jeddah, 21589 Saudi Arabia.
⁴ School of Computer Engineering, KIIT Deemed to University, Bhubaneswar, India.
⁵ Department of Infection Diseases, The Fourth People's Hospital of Huai'an, Huai'an, 223002 Jiangsu China.

^# Contributed equally.

Abstract

COVID-19 is a positive-sense single-stranded RNA virus caused by a strain of coronavirus, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Several noteworthy variants of SARS-CoV-2 were declared by WHO as Alpha, Beta, Gamma, Delta, and Omicron. Till 13/Dec/2022, it has caused 6.65 million death tolls, and over 649 million confirmed positive cases. Based on the convolutional neural network (CNN), this study first proposes a ten-layer CNN as the backbone model. Then, the exponential linear unit (ELU) is introduced to replace ReLU, and the traditional convolutional block is now transformed into conv-ELU. Finally, an ELU-based CNN (ELUCNN) model is proposed for COVID-19 diagnosis. Besides, the MDA strategy is used to enhance the size of the training set. We develop a mobile app integrating ELUCNN, and this web app is run on a client-server modeled structure. Ten runs of the tenfold cross-validation experiment show our model yields a sensitivity of $94.41 \pm 0.98$ , a specificity of $94.84 \pm 1.21$ , an accuracy of $94.62 \pm 0.96$ , and an F1 score of $94.61 \pm 0.95$ . The ELUCNN model and mobile app are effective in COVID-19 diagnosis and give better results than 14 state-of-the-art COVID-19 diagnosis models concerning accuracy.

Keywords: COVID-19; Cloud computing; Convolutional neural network; Cross validation; Deep learning; Exponential linear unit; Mobile app; Multiple-way data augmentation; SARS-CoV-2.