COVID-19 Diagnosis Based on Deep CNN using CT-Scans
Abstract
On March 11, COVID-19 becomes a worldwide pandemic. Rapidly spreading rate nature of Covid-19 as well as no specific drugs for Covid-19 are available yet, makes it necessary to detect and quarantine the infected people in an early stage as much as possible to decrease COVID-19 separation. The approach of testing COVID-19 is by using (RT-PCR) which suffers from low sensitivity and time-consuming issues. Therefore, chest CT-scans can be used for diagnostic purposes due to their sensitivity comparing to RT-PCR. In this paper, we built a deep CNN algorithm to detect COVID-19 utilizing CT-scans. CT-CNN deep learning architecture used for accurate detection of COVID-19. The results of our proposed model are very promising, with an accuracy of 96.14%, F1 score of 96.21%, and Recall 97.53% on classifying CT-scans into infected by covid-19 or not.