COVID-19, an upper respiratory viral infection, is spreading too quickly and is now found almost everywhere on the planet. A chest x-ray and RT-PCR are used to confirm Covid-19. However, resource constraints in developing countries make this difficult. This paper aims to monitor infection in Egyptian's lungs using chest X-rays and convolutional neural networks in the early-stage detection. The omicron chest X-ray dataset is gathered from public records along with hospital and doctor arrangements that are approved by their patients to achieve our goal. The classification mission's backbone is a classic neural network (CNN), which benefited from its speed while deep networks are being trained and reduced the impact of vanishing gradient problems. Images are resized and pre-processed to improve sharpness and contrast before being used to validate the proposed system. Meanwhile, by feeding images into a deep neural network, infectiousness can be predicted. The deep learning measurement's receiver operating characteristics (ROC) region under the curve is 0.9888, with 96.2 percent sensitivity, 98 percent accuracy, and 100 percent precision, according to the findings. It has been demonstrated that the proposed system can be easily modified to increase efficiency by adding additional images (normal and infected). As a quick alternative to the current PCR-based method, the proposed system offers a significant advantage to all nations in terms of screening and diagnosing OMICRON. |
