Semi-automatic Segmentation of COVID-19 Infection in Lung CT Scans

As the prevalence of COVID-19, concerns about the treatment of the disease and its impact on communities’ future have increased sharply. The best way to prevent the spread of COVID-19 disease is to quickly diagnose patients and prevent them from coming into contact with healthy people. Computer methods are very effective in finding patients with COVID-19 and speed up the diagnosis. These methods are also widely used to assess a patient’s condition, for example, to assess the disease’s progression over time and to measure the rate of spread of the virus in the lungs. In this article, a segmentation method is introduced to segment the infected parts of the lung in CT scans. This method is based on Lazy-Snipping and Super-pixel algorithms. As a result of segmentation, the performance of algorithm is presented and compared with other methods using Dice score which was 80%.