Efficient framework for detecting COVID-19 and pneumonia from chest X-ray using deep convolutional network

Recently, the COVID-19 pandemic is considered the most severe infectious disease because of its rapid spreading. Radiologists still lack sufficient knowledge and experience for accurate and fast detecting COVID-19. What exacerbates things is the significant overlap between Pneumonia symptoms and COVID-19, which confuses the radiologists. It's widely agreed that the early detection of the infected patient increases his likelihood of recovery. Chest X-ray images are considered the cheapest radiology images, and their devices are available widely. This study introduces an effective Deep Convolutional Neural Network (DCNN) called "DeepChest" for fast and accurate detection for both COVID-19 and Pneumonia in chest X-ray images. "DeepChest" runs with a small number of convolutional layers, a small number of max-pooling layers, and a small number of training iterations compared with the recent approaches and the state-of-the-art of DCNN. We conducted the experimental evaluations of the proposed approach on a data set with 7512 chest X-ray images. The proposed approach achieves an accuracy of 96.56% overall, 99.40% in detecting COVID-19, and 99.32% in detecting Pneumonia. In actual practice, the presented approach can be used as a computer-aided diagnosis tool to get accurate results in detecting Pneumonia and COVID-19 in chest X-ray images. (C) 2022 THE AUTHORS. Published by Elsevier B.V. on behalf of Faculty of Computers and Information, Cairo University.