ABSTRACT
COVID-19 is caused by coronavirus-2 (SARS-CoV-2) virus strain. The disease is easily communicable and fatal in nature. As it is contagious, its early and rapid detection is crucial to lower the death rate. In our research, we explore the use of reconstructed image features based on graph-based techniques for classifying the CT-scan images of Covid-19 suspected patients for detection of the infection. We also compare our proposed approach with the traditional approach of using images directly for classification with convolutional neural networks. Results show that our proposed approach using reconstructed features yields 99% accuracy along with high sensitivity.
