Comparative Study on Detection of COVID-19 using different Convolution Layers

Abstract

The Coronavirus disease 2019 (COVID-19) is a highly infectious disease caused by the severe acute respiratory syndrome. The disease was first instigated in December 2019 in the place called Wuhan which is the capital of a province in China named Hubei and meanwhile, it has spread universally throughout the world. The impact is greatly influenced so that World Health Organization (WHO) has declared the ongoing pandemic of COVID-19 a Public Health Emergency. Artificial intelligence renders its support to analyze chest X-ray (CXR) images for COVID-19 diagnosis. This proposed system is aimed to automatically detect COVID-19 patients using digital chest x- ray images while increasing the accuracy of the model tried with different convolution layers. The dataset was created as a mixture of publicly available X-ray images from patients with confirmed COVID-19 disease and healthy folks. To alleviate the small number of samples, we have inculcated many data augmentation techniques that further enhance the accuracy of the model with different convolution layers. The research aims to design a Deep Learning based model for Covid 19 prediction through X-Ray images. The parameters chosen are applied over 3 different models designed by varying Convolution Layers and proved that accuracy enhances when number of layers increases.