Performance Enhancement of Covid-19 Chest X-Ray Image Classification Using GAN and CNN

Abstract

Covid-19, a respiratory disease caused by the SARS-CoV-2 virus, manifests in individuals with varying degrees of severity. Chest X-rays serve as initial screening procedures for suspected Covid-19 infections, aiding in the detection of abnormalities. Various approaches utilizing deep learning models like Convolutional Neural Networks (CNN) have been proposed for Covid-19 detection through the analysis of chest radiograph images. However, the availability of radiographic Chest X-ray images for Covid-19 remains limited and not easily accessible to the global research community. This scarcity of data poses a significant challenge for further research in the diagnosis of Covid-19 using radiographic images. This research aims to overcome this obstacle by the generation of synthetic Chest X-ray images through the use of Generative Adversarial Networks (GANs) which produce more realistic images compared to traditional data augmentation methods like rotation, scaling, and flipping. The research analyzes and compares the results to address the limited availability of Covid-19 Chest X-ray data. By augmenting the existing dataset, this research has aimed to improve the performance of Covid-19 diagnosis models based on past research findings.