Welcome to the InfoSci Platform
Fine Tuning CNN for COVID-19 Patterns Detection From Chest Radiographs

Fine Tuning CNN for COVID-19 Patterns Detection From Chest Radiographs

Anju Jain, Saroj Ratnoo, Dinesh Kumar
Copyright: © 2022 |Volume: 11 |Issue: 4 |Pages: 15
ISSN: 2160-9551|EISSN: 2160-956X|EISBN13: 9781683182603|DOI: 10.4018/IJRQEH.308801
Cite Article Cite Article

MLA

Jain, Anju, et al. "Fine Tuning CNN for COVID-19 Patterns Detection From Chest Radiographs." IJRQEH vol.11, no.4 2022: pp.1-15. http://doi.org/10.4018/IJRQEH.308801

APA

Jain, A., Ratnoo, S., & Kumar, D. (2022). Fine Tuning CNN for COVID-19 Patterns Detection From Chest Radiographs. International Journal of Reliable and Quality E-Healthcare (IJRQEH), 11(4), 1-15. http://doi.org/10.4018/IJRQEH.308801

Chicago

Jain, Anju, Saroj Ratnoo, and Dinesh Kumar. "Fine Tuning CNN for COVID-19 Patterns Detection From Chest Radiographs," International Journal of Reliable and Quality E-Healthcare (IJRQEH) 11, no.4: 1-15. http://doi.org/10.4018/IJRQEH.308801

Export Reference

Mendeley
Favorite Full-Issue Download

Abstract

The COVID-19 pandemic has crumbled health systems all over the world. Quick and accurate detection of coronavirus infection plays an important role in timely referral of physicians and control transmission of the disease. RT-PCR is the most widely test used for identification of COVID-19 patients, but it takes long to deliver the report. Researchers around the world are looking for alternative machine learning techniques including deep learning to assist the medical experts for early COVID-19 disease diagnosis from medical imaging such as chest films. This study proposes an enhanced convolutional neural network (EConvNet) model for the presence and absence of coronavirus disease from chest radiographs to contain this pandemic. The model is accurate compared to the traditional machine learning algorithms (RF, SVM, etc.). The suggested CNN model is approximately as accurate as the classifiers based on transfer learning (such as InceptionV3, VGG16, and Densenet121). Despite being simple in terms of number of parameters learnt, it takes less training time and demands less memory.