Performance Analysis of Deep Learning Frameworks for COVID 19 Detection

Abstract

The Coronavirus known as COVID-19 is one of the biggest pandemic in the human history. About 3.12 million deaths and more than 150 million cases have been diagnosed to date. The economic cost of the virus is also huge, due to lockdown in many parts of the world. The traditional method to detect COVID-19 is a PCR test, but it takes about 4–5 hours to get the results. Secondly, in some cases, the false-negative ratio is high in PCR kits. As an alternative method, radiology images, like CT-scan and chest X-rays, can be used for COVID-19 diagnosis. Deep learning has shown remarkable results in medical image analysis, such as tumor segmentation. This paper evaluates four famous CNN models, VGG-16, AlexNet, ResNet18, and Inception V3, on the dataset complied with the chest X-ray images for COVID-19 patient. The models are trained on two public datasets. The simulation results show the effectiveness of deep learning models for COVID-19 detection.