Deep Learning Algorithms for Diagnosing Covid 19 Based on X-Ray and CT Images

Abstract

An outbreak of a highly pathogenic coronavirus, which can cause chronic respiratory illness and high mortality rates. It takes a considerable amount of time to perform the polymerase chain reaction (PCR) used in COVID tests. Its accuracy ranges from 30% to 70%. In contrast, CT and chest X-ray diagnostics are 98% and 80% accurate in detecting COVID, respectively. A deep learning algorithms was applied to CT and X-ray images to enable rapid and accurately diagnosis of COVID-19 within seconds. In this survey, we revised all state-of-the-art studies of COVID-19 based on CT and X-ray images. Also, we analysed multiple deep learning networks and compared the performance of each technique. The result of the comparison shows that the baseline neural network has better efficiency in the recognition of COVID-19. The detection accuracy of baseline networks ranges between 93% and 98.7%. This shows the efficiency of deep learning techniques in identifying COVID-19.