Identification of Corona Virus From X-ray Images

Abstract

In late December 2019, China announced the novel COVID-19 coronavirus. And after that, it spread all around the world, taking the place of a pandemic. In the meantime, there are many testing kits available but it takes time to get test results so we need to fast disease detection. for that reason, there is some work done on neural network models. all those images are performed on CXR data that are publicly available. In this study, we also work on five different pre-trained models(resnet-50, Inception V3, Inception Resnet V2, Resnet-101, Resnet-152) for the classification and detection of these diseases and we also have compared their accuracy on different epoch values. After evoluted all the models throughout the accuracy, recall, f1-scores we have got that resnet-50 perform 100% accuracy, Recall 100% and Precision 100% with 30 epoch value and also Resnet-101 perform 100% accuracy on epoch value 40. Those two models have given better results to compare with the other models.