An efficient Automatic COVID-19 Detection from Medical Images using hybrid CNN-RNN and GAN network

Coronavirus, commonly known as COVID-19, is a viral illness caused by the SARS-CoV-2 virus, which stands for severe acute respiratory syndrome coronavirus 2. The global spread of COVID-19 has had a deleterious effect on both the public health and economy. One major step in the battle versus COVID-19 is detecting the virus in patients through positive chest X-rays. Early studies have identified abnormalities in the chest X-rays of infected patients that are indicative of the disease. Research has shown high accuracy in identifying COVID-19 patients using chest X-rays, which has spurred the development of various deep learning algorithms. Convolutional neural networks (CNNs), a type of deep learning model, require large amounts of training data. However, due to the recent emergence of the pandemic, gathering a substantial dataset of radiographic images in a short time has been difficult. To address this, our study introduces a model called CovidGAN, which employs a Generative Adversarial Network (GAN) to generate synthetic chest X-ray (CXR) images. Additionally, we propose a hybrid CNN-RNN network for identifying COVID-19 in X-ray images, achieving a classification accuracy of 98.75%.