Lisnet: A Covid-19 Lung Infection Segmentation Network Based on Edge Supervision and Multi-Scale Context Aggregation

Abstract

Corona Virus Disease 2019 (COVID-19) spread globally in early 2020, leading to a new health crisis. Automatic segmentation of lung infections from computed tomography (CT) images provides an important basis for early diagnosis of COVID-19 quickly. In this paper, we propose an effective COVID-19 Lung Infection Segmentation Network (LISNet) based on edge supervision and multi-scale context aggregation. More specifically, an Edge Supervision module is introduced to the feature extraction part to enhance the low contrast between lesions and normal tissues. In addition, the Multi-scale Feature Fusion module is added to enhance the segmentation ability of different scales Lesions. Finally, the Context Aggregation module is used to aggregate high- and low-level features and generate global information. Experiments demonstrate that our method outperforms other state-of-the-art methods on the public COVID-19 CT segmentation dataset.