COVID-19 Prediction using X-Ray Images and Deep Learning Model

Abstract

In the year 2019, the novel corona virus was emerged in China. Since, then every individual has put their own self to fight against this pandemic situation. Till April 14, 2022, the total COVID-19 confirmed cases are 500,186,525 and deaths are 6,190,349 worldwide. RT-PCR and Rapid antigen test are the standards for COVID-19 detection by WHO. These both tests are time consuming and costly and also there are higher possibilities of false results. On this conquest, we decided to implement a CNN model to detect COVID-19 patients from chest X-ray images. Convolutional Neural Networks (CNN) are important in the field of deep learning for diagnosing a variety of diseases, such as Coronary Artery Disease, Alzheimer’s disease, and Parkinson’s disease. Similarly, CNN has a good chance of identifying COVID-19 patients using diagnostic pictures such chest X-rays. The proposed model performs with a 97.84% accuracy and 99.11% precision. Similarly, the Receiver Operating Characteristic (ROC) has the AUC of 97.8% and the F1-score generated is 97.81. All these results can be improved by using different techniques and equally larger dataset for training of model.