Automated Detection of COVID-19 from X-ray Images using Deep Convolutional Neural Networks

Abstract

Coronavirus, also known as COVID19, is a dangerous disease that has put many people's lives in jeopardy around the world by damaging the lungs directly. The detection of coronavirus is a challenging medical procedure due to its increasing cases. Currently, the use of x-ray images for coronavirus diagnosis is commonly used. Recently, various deep learning based models have been used for image classification. These models have generated competitive results in terms of feature selection and classification. In this article, we proposed a set of seven pretrained neural network models (VGG16, VGG19, InceptionV3, ResNet50, Xception, DenseNet121 and InceptionResNetV2) for the detection of coronavirus infection using chest X-ray images collected from an open source. It was observed that out of these models, pretrained DenseNet121 yielded highest classification accuracy of 97% for the particular dataset.