الفهرس | Only 14 pages are availabe for public view |
Abstract This thesis studies deep learning models for analyzing COVID-19 disease. These models aid in distinguishing patients’ cases such as COVID-19, viral pneumonia, lung opacity, or healthier with high performance. Deep learning approaches are strongly encouraged in developing these expert COVID-19 models, which can aid clinicians in the early diagnosis and prediction of COVID-19 disease. Deep learning models are trained using a neural network architecture or a set of labeled data that contains multiple layers. These architectures learn features directly from the data without hindrance to manual feature extraction [1]. In this thesis, the performance of several COVID-19 imaging approaches is examined including Chest X-ray (CXR), and Computed Tomography (CT). Medical CT scans and CXR images are used as inputs to classification methods that emphasize the role of physiological changes unique to COVID-19. For the analysis of COVID-19 in medical images, this thesis studies a computer aided design model that can recognize positive COVID-19 and other lung diseases cases. These demonstrated the pipeline of medical images and estimation procedures involved in COVID-19 image acquirement and diagnosis, utilizing Computed Tomography (CT) scans and Chest X-Rays (CXR) images[2]. This thesis studies two deep learning models, namely AlexNet and VGG-16, to detect cases of COVID-19. These classification approaches are applied to several medical images to detect COVID-19 and other lung diseases. The two proposed architectures have achieved a multi-classes classification accuracy of 97.28% and 98.8%, respectively, for CXR images, while achieving 98.41% and 99.42% accuracy for CT images. Finally, the results of the two deep models showed better performance compared to other recent Studies. |