الفهرس | Only 14 pages are availabe for public view |
Abstract Corona Virus Disease 19 (COVID-19) is a hazardous disease that has endangered the health of many people around the world by directly affecting the lungs. It firstly spread in China since December 2019. Then, it has spread at a high rate around the world. Therefore, rapid diagnosis of COVID-19 has become a very hot research topic. As COVID-19 pandemic has already affected the world like no other pandemic disease in the history and still, the people are dealing with this deadly pandemic situation. Its diagnosis as well as prognosis has eventually become a huge challenge for the medical fraternity. Artificial Intelligence (AI) methodologies can be used to obtain consistent high performance for diagnosing COVID-19. Among the AI methodologies, Deep Learning (DL) networks have gained much popularity compared to traditional Machine Learning (ML) methods. Chest X-ray and Computed Tomography (CT) scan are the most important medical imaging techniques for diagnosing COVID-19. All researchers are looking for effective solutions and fast treatment methods for this epidemic. To reduce the need for medical experts, fast and accurate automated detection techniques are introduced. Deep Convolutional Neural Network (DL-CNN) is a special type of neural networks, which can automatically learn representations from the data. The DL-CNN technologies are showing remarkable results for detecting cases of COVID-19. Our thesis presents different DL-CNN models for COVID-19 detection for different medical imaging modalities like CT and X-ray. Efficient CNN-based model using a wireless communication and classification system is presented, which achieves accuracy of 98.4% for classification of COVID-19 cases. Furthermore, the proposed Deep Feature Concatenation (DFC) mechanism and CNN-based fusion technique are suggested. The DFC algorithm provides higher accuracy of 99.3% for classification Xray images. The fusion of X-ray and CT features gives an accuracy results of 99.01% for augmented dataset, and 96.5% for non-augmented dataset. |