CNN Based Analysis of COVID-19 Using Chest X-Ray Images

Abstract

Coronavirus (COVID – 19) is a deadly virus that originally originated from China’s Wuhan district around November last year. It has a deadly effect on the human respiratory system if the condition escalates. Currently, millions of people have been affected worldwide, and in countries like India, the cases are still on the rise. Due to an increased rise in cases, the testing facilities are struggling to keep up with the demand for testing, and medical experts are looking for alternate ways to speed up testing. In this paper, we have experimented with one such way by developing a CNN-based model to classify the chest X-ray images for the detection of coronavirus affected cases. For result analysis, we have applied CNN based VGG 16, VGG 19, and custom model. Further, we compare the result of these models based on accuracy. In this experimental analysis, VGG 19 model detected 99% of COVID-19 infected cases accurately as compared to VGG 16. This is entirely an experimental study and should not be used in real-life scenarios without an evaluation by medical experts and determining the effectiveness of this method.