Detection of Covid-19 in Chest X-Ray Images using Deep Learning

Abstract

: COVID-19 is a new kind of virus that was first emerged in China in December 2019 and now has 14,29,65,972 confirmed cases worldwide. Detection is a primary part of curb the coronavirus's spread, mainly based on test results such as PCR, RTPCR, high temperature, X-Ray, and CT-Scans, and Covid-19 symptoms. The research on detecting the coronavirus has gained momentum; scientists and doctors are working to find new techniques to detect the virus with higher accuracy and less time. Deep Learning and Transfer Learning play an essential role in the detection and classification of various abnormalities in medical image datasets with state-of-the-art convolutional neural networks. This paper proposes using seven transfer learning models to classify X-ray images into three classes: Normal, COVID-19, and Viral Pneumonia using a dataset of 15,189 X-ray images. We have used VGG-16, VGG-19, MobileNet-V2, DenseNet-121, Xception, ResNet-50V2, and Inception-V3 models. As a result, we achieved the best classification accuracy of 99.0%, 98.2%, 97.8%, 97.6%, 97.4%, 97.0%, and 95.3% respectively. It is assumed that the doctors and medical staff could use the paper's methods for diagnosing the COVID-19 disease with higher accuracy.