Detection of Coronavirus illness using Techniques of Deep Learning and CNN

Abstract

A year has been passed with the global pandemic creating havocs in everyone’s life. The novel Coronavirus is still raging around the globe causing catastrophic consequences on the entire health and wealth of humankind. Tests are being conducted in an insane amount on the suspected individuals. Infections that are gained through respiratory course, for example, the lethal SARS-CoV-2, are determined to have the assistance of direct identification of viral parts in respiratory examples. The two most generally utilized techniques to do this are nucleic corrosive enhancement tests through polymerase chain response/reaction (PCR) or antigen-based tests. This can take a while to generate results as there is steady increase in number of cases and causing delay in laboratories. Early detection of the virus is life saviour, if the virus is left unnoticed it can be fatal for ones’ life. The current industrial era is ruled by fields of artificial intelligence and machine learning; hence this paper is an attempt to use one of these practices for novel corona virus prediction using chest radiogram images. Here dataset of Chest Roentgenogram images of patients infected with the corona virus and normal Chest Roentgenogram images are used to detect coronavirus infection. The study employs an efficient approach of application Convolutional Neural Network in predicting if the patient is affected and unaffected with the virus. The prepared model created a precision pace of 92.77% at the time of the performance preparation.