Detection of COVID-19 from radiology modalities and identification of prognosis patterns

Document Type : Research Paper


1 Department of IT, V R Siddhartha Engineering College, India

2 College of Medicine, University of Anbar, Ramadi, Iraq

3 Department of Nursing Sciences, Shirvan Branch, Islamic Azad University, Shirvan, Iran

4 Department of Computer Science, College of Computer Science and Information Technology, University of Anbar, Ramadi, Iraq

5 Department of Electricity Engineering College of Engineering, University of Tikrit, Tikrit, Iraq

6 Department of Communication Engineering Collage of Engineering, University of Diyala: Baqubah, Diyala, Iraq

7 Directorate of Private University Education, Ministry of Higher Education, Baghdad, Iraq

8 Department of Computer Engineering Collage of Engineering, University of Diyala: Baqubah, Diyala, Iraq


SARS-CoV-2 and the consequential COVID-19 virus is one of the major concerns of the 21st century. Pertaining to the novelty of the disease, it became necessary to discover the efficacy of deep learning techniques in the quick and consistent discovery of COVID-19 based on chest X-ray and CT scan image analysis. In this related work, Prognostic tool using regression was designed for patients with COVID-19 and recognizing prediction patterns to make available important prognostic information on mortality or severity in COVID-19 patients. And reliable convolutional neural network (CNN) architecture models (DenseNet, VGG16, ResNet, Inception Net)to institute whether it would work preeminent in terms of accuracy as well as efficiency with image datasets with Transfer Learning. CNN with Transfer Learning were functional to accomplish the involuntary recognition of COVID-19 from numerary chest X-ray and CT scan images. The experimental results emphasize that selected models, which is formerly broadly tuned through suitable parameters, executes in extensive levels of COVID-19 discovery against pneumonia or normal or lung opacity through the precision of up to 87% for X-Ray and 91% intended for CT scans.


Volume 13, Issue 1
February 2022
Pages 1351-1365
  • Receive Date: 29 October 2021
  • Accept Date: 29 October 2021
  • First Publish Date: 29 October 2021