Multi-objective Genetic Algorithm Based Deep Learning Model for Automated COVID-19 Detection Using Medical Image Data.

In early 2020, the world is amid a significant pandemic due to the novel coronavirus disease outbreak, commonly called the COVID-19. Coronavirus is a lung infection disease caused by the Severe Acute Respiratory Syndrome Coronavirus 2 virus (SARS-CoV-2). Because of its high transmission rate, it is crucial to detect cases as soon as possible to effectively control the spread of this pandemic and treat patients in the early stages. RT-PCR-based kits are the current standard kits used for COVID-19 diagnosis, but these tests take much time despite their high precision. A faster automated diagnostic tool is required for the effective screening of COVID-19. In this study, a new semi-supervised feature learning technique is proposed to screen COVID-19 patients using chest CT scans. The model proposed in this study uses a three-step architecture, consisting of a convolutional autoencoder based unsupervised feature extractor, a multi-objective genetic algorithm (MOGA) based feature selector, and a Bagging Ensemble of support vector machines based binary classifier. The proposed architecture has been designed to provide precise and robust diagnostics for binary classification (COVID vs.nonCOVID). A dataset of 1252 COVID-19 CT scan images, collected from 60 patients, has been used to train and evaluate the model. The best performing classifier within 127 ms per image achieved an accuracy of 98.79%, the precision of 98.47%, area under curve of 0.998, and an F1 score of 98.85% on 497 test images. The proposed model outperforms the current state of the art COVID-19 diagnostic techniques in terms of speed and accuracy. The experimental results prove the superiority of the proposed methodology in comparison to existing methods.The study also comprehensively compares various feature selection techniques and highlights the importance of feature selection in medical image data problems.