COVID-19 Diagnosis Applied DWT and CNN on X-ray Chest Images

Abstract

Background: Medical images have many important applications, and this importance increased when the emergence of the COVID-19 pandemic. These applications have been focused on computed tomography chest images and X-ray images. This research will focus on special X-ray medical image applications of coronavirus (COVID-19). Methods: Many methods are applied on medical images to achieve certain features. The designed approach is implemented through many steps starting from preprocessing up to classification step. The proposed approach focusing on generating efficient features using discrete wavelet transform (DWT) then applying convolutional neural network (CNN) to classify between normal and abnormal COVID-19. Results: The COVID-19 diagnosis approach is implemented to achieve high performance system. The obtained result of COVID-19 diagnosis applied CNN tool leading to validation accuracy of 92.31%. Conclusion: Hybridizing two technologies (DWT and CNN) is intended to reach the best results in the diagnostic process. In addition, X-ray chest image is an important tool for detection and diagnosis of COVID-19 diseases.

Keywords

• coronavirus
• diagnosis
• chest images
• discrete wavelet transform
• convolutional neural network