Automatic X-ray COVID-19 Lung Image Classification System based on Multi-Level Thresholding and Support Vector Machine

Aboul Ella Hassanien; Lamia Nabil Mahdy; Kadry Ali Ezzat; Haytham H. Elmousalami; Hassan Aboul Ella

U-Net++ neural network model has been developed for processing over 6000 CT scans to classify the cases of patients into COVID-19 infected or not infected [19] . The model produced 93.55% and 100% for specificity and sensitivity, respectively. In addition, the model presented negative positive value (NPV) of 100%, positive value (PPV) of 84.62%, and a total accuracy of 95.24%. The model greatly helps radiologists by decreasing the reading time of CT scans by 65% [19] . On the other hand, 3-category classification model is designed to distinguish the COVID-19 cases where both Song et al. (2020) and Xu et al. (2020) have applied feature extraction using Feature Pyramid Network and several fullyconnected layers for cases classifications [20, 21] . The feature extraction using Feature Pyramid Network model produces an acceptable total accuracy of 86.7% using CT scans. The . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.