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 existing methods and databases for Lung X-Ray, combining hand-crafted descriptors with deep networks, training a classification module, and validating the classification results using sensitivity and specificity parameters. The proposed network will have a different number of convolution layers and use a different optimizer, such as stochastic gradient descent sgd and Adam, for better convergence. The project will also design a user-friendly application interface, called COVID-19 TeevrataMaapak CTM, which will be used for initial screening and treatment. |