Executive Summary : | The project aims to develop a mathematical model to detect COVID-19 positive cases using chest X-ray images without considering symptoms. Existing methods focus on predicting positive cases but do not estimate the severity of the positive case. The proposed model aims to estimate the severity of COVID-19 in a patient, design a scalable CNN to increase detection accuracy, and reduce doctor intervention in the initial screening process. The project will be divided into four major tasks: identifying suitable methods and databases for Lung X-Ray, implementing hand-crafted descriptors, analyzing feature-level fusion of hand-crafted and deep features, and training a classification module. The proposed network will use conventional CNNs like ResNet50, VGG19, and ImageNet as the base model for COVID-19 detection from chest X-rays. The network will have a different number of convolution layers and use a different optimizer for detection and severity detection. The project aims to improve the accuracy of detection and reduce doctor intervention in the initial screening process. |