Identification of COVID-19 using chest X-Ray images

Abstract

ABSTRACT SARS-COV2 or the coronavirus disease even after two years of identification is still a problem because of its varying mutations. The prediction and classification of the virus are one of the most important tasks for controlling the spread of the pandemic. So immediate development of some strong techniques with artificial intelligence implementation to predict COVID-19 is required. The paper’s main aim is to develop an accurate, efficient, and time-saving model for detecting COVID-19 from Chest X-Ray (CXR) images. This study utilises a dataset from the Kaggle database containing CXR images of COVID-19, Viral Pneumonia, and Normal Healthy lung images. Here, some Deep Learning (DL) models like GoogleNet, SqueezeNet, and ResNet-18 are utilised for the detection of the virus. The overall accuracy achieved from the models like GoogleNet is 92.46%, for SqueezeNet is 92.57%, and ResNet-18 is 97.38%. From the obtained outcomes it is concluded that ResNet-18 provides accurate results as compared with the other three models. The achieved outcomes could help medical experts in the future to predict and classify COVID-19 and can very quickly diagnose cases with effective models.