Detection of COVID-19 Based on Chest X-rays Using Deep Learning
Walaa Goudaa✉, Maram Almurafehb, Mamoona Humayunb, Noor Zaman Jhanjhic d
- aDepartment of Computer Engineering and Network, College of Computer and Information Sciences, Jouf University, Sakaka 72341, Aljouf, Saudi Arabia
- bDepartment of Information Systems, College of Computer and Information Sciences, Jouf University, Sakaka 72341, Aljouf, Saudi Arabia
- cSchool of Computer Science and Engineering (SCE), Taylor’s University, Subang Jaya 47500, Selangor, Malaysia
- dSchool of Computer Science and Engineering (SCE), Taylor's University, Subang Jaya 47500, Selangor, Malaysia
- ✉Corresponding author
Papers open inside SciRah from the original publisher or repository; SciRah never hosts them.
Abstract
The coronavirus disease (COVID-19) is rapidly spreading around the world. Early diagnosis and isolation of COVID-19 patients has proven crucial in slowing the disease's spread. One of the best options for detecting COVID-19 reliably and easily is to use deep learning (DL) strategies. Two different DL approaches based on a pertained neural network model (ResNet-50) for COVID-19 detection using chest X-ray (CXR) images are proposed in this study. Augmenting, enhancing, normalizing, and resizing CXR images to a fixed size are all part of the preprocessing stage. This research proposes a DL method for classifying CXR images based on an ensemble employing multiple runs of a modified version of the Resnet-50. The proposed system is evaluated against two publicly available benchmark datasets that are frequently used by several researchers: COVID-19 Image Data Collection (IDC) and CXR Images (Pneumonia). The proposed system validates its dominance over existing methods such as VGG or Densnet, with values exceeding 99.63% in many metrics, such as accuracy, precision, recall, F1-score, and Area under the curve (AUC), based on the performance results obtained.
TL;DRTwo different DL approaches based on a pertained neural network model (ResNet-50) for COVID-19 detection using chest X-ray (CXR) images are proposed in this study, with values exceeding 99.63% in many metrics. (Semantic Scholar)