Convolution Neural Network Architectures for COVID-19 Detection using Images from Chest X-Rays: A Review

Actually, the unanticipated Crisis could not have been predicted—has been the new Coronavirus. Since a large number of individuals need to be screened, it has become a pandemic & generated concerns about the state of the healthcare system and facilities. The prevalent diagnostic procedure is RT-PCR. However, there are a number of problems with RTPCR testing include relying solely on one method even though it is becoming more and more challenging for many nations to obtain the necessary number of test samples. Additionally, there have been several instances of false positive findings .All of these considerations make it necessary to look into various testing procedures; it is not sufficient to rely solely on one way. Chest X-rays are a wonderful tool to learn more about status & seriousness of CV19in addition to RT-PCR tests. When one person’s chest X-rays have been obtained, a decision regarding their CV19state must be made. This requires a high degree of precision & experience, both of which are again in low supply. Therefore, a CV19detecting system has been designed as a potential answer to this issue. This method can help the medical specialists in their analysis of the report before they reach a decision.Given that our goal is analogous to medical picture classification and that Deep Learning methods, particularly Convolutional Neural Networks (CNN), have demonstrated their superior performance in this area, CNN is a strong fit for our application case. In order to diagnose Covid-19, we examined 4 different CNN architectures using chest X-ray. Pre-training of the models is done on Image Net datasets. Results have been produced by using transfer learning. A comparison of the outcomes from various architectural approaches demonstrates that system outcomes on CNN offer excellent promise for CV19detection and diagnosis.