A comparative study of three deep learning-based methods for segmentation of lung lesions associated with COVID-19 in 3D CT images

Abstract

In this paper, we present three different methods for segmentation of lung lesions associated with COVID19 virus disease in 3D CT images. All the methods are based on using state-of-the art deep learning techniques. The proposed methods were tested in the framework of COVID-19 Lung CT Lesion Segmentation Challenge 2020 (COVID-19-20), a MICCAI-associated event which was held on the «Grand Challenge» platform. The best achieved result corresponded to the 6th rank among 98 challenge finalists from all over the world. The results of this study demonstrate the importance of development of robust image analysis solutions applicable to the image data acquired in different hospitals.