DenseNet-121 Model for Diagnosis of COVID-19 Using Nearest Neighbour Interpolation and Adam Optimizer

The pandemic of COVID-19 has caused disaster all over the world since its outbreak. It is a disease which affects the respiratory system of patient affecting the lung area, which then causes severe infection, that can lead to death of the patient. Researchers use medical images for its initial diagnosis. In this paper, we apply DenseNet-121 model for classifying chest X-ray images of patients into COVID-19. We use nearest neighbour interpolation method for data preprocessing and Adam optimizer for faster convergence. Experiments have shown that our methodology of using DenseNet-121 with nearest neighbour interpolation technique shows improved accuracy than state-of-the-art deep learning architectures. We compare the results obtained by using DenseNet-121 model with various interpolation methods. We achieve higher accuracy of 95.27 % by using nearest neighbour interpolation method for DenseNet-121 as compared to the other interpolation methods.