Attention Residual Convolution Neural Network Based on U-Net for COVID-19 Lung Infection Segmentation.

In 2021, the COVID-19 is still widespread around the world, which has a great impact on peopleʼs daily lives. However, there is still a lack of research on the fast segmentation of lung infections caused by COVID-19. The segmentation of the COVID-19- infected region from the lung CT is of great significance for the diagnosis and care of patients. In this paper, attention gate residual U-Net (AGRU-Net) based on residual network and attention gates is proposed for the segmentation. As COVID-19- infected regions varies greatly from one to another, the deeper network is needed to extract segmentation features. The residual unit is an effective solution to the degradation problem of deeper network. The addition of attention gates to U-Net suppresses irrelevant areas in the image for more significant segmentation characteristics. In this paper, the experiments on a public COVID-19CT dataset show that AGRU-Net has good performance in the segmentation of COVID-19- infected region.