SVM Based COVID-19 Detection from CT Scan Image using Local Feature

During the outbreak time of pandemic COVID-19, we are so much worried and scared about our life. Going to the hospital is a very risky job to test COVID-19 because ordinary people can be easily infected by the COVID-19 patient. And also it is very time-consuming to get the test result of the COVID-19. In order to focus on these issues, a machine learning model is designed to detect COVID-19 using the local feature of CT-scan images and Support Vector Machine (SVM). A publicly available SARS-CoV-2 (severe Acute Respiratory Syndrome Coronavirus 2) dataset is used in the proposed system which contains 2481 CT scan images in total where 1252 samples are COVID and the rest of the samples are Non-COVID. Local Binary Pattern (LBP) algorithm is applied to extract local features from images. We also compare our proposed model with some state-of-the-art algorithms such as Decision Tree (DT), K-Nearest Neighbor (KNN), and AdaBoost based on some evaluation matrices i.e. precision, recall, and F1 score. According to the result, the proposed model provides 97.09% F1 score which is better to detect COVID-19 for clinical usage.