14 Deep convolutional neural network for the classification of COVID-19 from chest X-ray images

Abstract

Coronavirus disease (COVID-19) caused by a new member of the coronavirus family is a respiratory disease that has quickly reached the level of a pandemic with high morbidity and mortality. In just a few months, it had a huge impact on the society and the world economy. COVID-19 presents many challenges to all aspects of healthcare, including reliable approaches of analysis, treatment, and prevention. Deep learning is one of the most effective artificial intelligence techniques for investigating chest X-ray images for effective and reliable COVID-19 screening. A vital step in the combat toward COVID-19 is a successful screening of contaminated patients, with one of the key screening approaches being radiological imaging using chest radiography. In this study, machine learning classification techniques were utilized to classify whether images represent COVID-19, viral pneumonia, and normal chest X-ray images. Because accuracy is the most significant feature in this issue, the accuracy of the deep convolutional neural network (DCNN) is improved by taking a large number of images to train the system and to increase the amount of iterations. This study aimed to automatically detect COVID‐19 pneumonia patients using digital chest X‐ray images while maximizing the accuracy in detection using DCNN. The dataset consists of 864 COVID‐19, 1,345 viral pneumonia, and 1,341 normal chest X‐ray images. In this study, DCNN-based model Inception V3 with transfer learning has been proposed for the detection of coronavirus pneumonia-infected patients using chest X-ray radiographs and achieved more than 98% of classification accuracy (training accuracy of 97% and validation accuracy of 93%). The results demonstrate that transfer learning was proved to be effective, and showed robust performance and easily deployable approach for COVID-19 detection. Meanwhile, the high accuracy of this computer-aided diagnostic tool can significantly improve the speed and accuracy of COVID-19 diagnosis.