Toward data-efficient learning: A benchmark for COVID-19 CT lung and infection segmentation.

PURPOSE: Accurate segmentation of lung and infection in COVID-19 computed tomography (CT) scans plays an important role in the quantitative management of patients. Most of the existing studies are based on large and private annotated datasets that are impractical to obtain from a single institution, especially when radiologists are busy fighting the coronavirus disease. Furthermore, it is hard to compare current COVID-19 CT segmentation methods as they are developed on different datasets, trained in different settings, and evaluated with different metrics. METHODS: To promote the development of data-efficient deep learning methods, in this paper, we built three benchmarks for lung and infection segmentation based on 70 annotated COVID-19 cases, which contain current active research areas, for example, few-shot learning, domain generalization, and knowledge transfer. For a fair comparison among different segmentation methods, we also provide standard training, validation and testing splits, evaluation metrics and, the corresponding code. RESULTS: Based on the state-of-the-art network, we provide more than 40 pretrained baseline models, which not only serve as out-of-the-box segmentation tools but also save computational time for researchers who are interested in COVID-19 lung and infection segmentation. We achieve average dice similarity coefficient (DSC) scores of 97.3%, 97.7%, and 67.3% and average normalized surface dice (NSD) scores of 90.6%, 91.4%, and 70.0% for left lung, right lung, and infection, respectively. CONCLUSIONS: To the best of our knowledge, this work presents the first data-efficient learning benchmark for medical image segmentation, and the largest number of pretrained models up to now. All these resources are publicly available, and our work lays the foundation for promoting the development of deep learning methods for efficient COVID-19 CT segmentation with limited data.

Active contour regularized semi-supervised learning for COVID-19 CT infection segmentation with limited annotations.

Infection segmentation on chest CT plays an important role in the quantitative analysis of COVID-19. Developing automatic segmentation tools in a short period with limited labelled images has become an urgent need. Pseudo label-based semi-supervised method is a promising way to leverage unlabelled data to improve segmentation performance. Existing methods usually obtain pseudo labels by first training a network with limited labelled images and then inferring unlabelled images. However, these methods may generate obviously inaccurate labels and degrade the subsequent training process. To address these challenges, in this paper, an active contour regularized semi-supervised learning framework was proposed to automatically segment infections with few labelled images. The active contour regularization was realized by the region-scalable fitting (RSF) model which is embedded to the loss function of the network to regularize and refine the pseudo labels of the unlabelled images. We further designed a splitting method to separately optimize the RSF regularization term and the segmentation loss term with iterative convolution-thresholding method and stochastic gradient descent, respectively, which enable fast optimization of each term. Furthermore, we built a statistical atlas to show the infection spatial distribution. Extensive experiments on a small public dataset and a large scale dataset showed that the proposed method outperforms state-of-the-art methods with up to 5% in dice similarity coefficient and normalized surface dice, 10% in relative absolute volume difference and 8 mm in 95% Hausdorff distance. Moreover, we observed that the infections tend to occur at the dorsal subpleural lung and posterior basal segments that are not mentioned in current radiology reports and are meaningful to advance our understanding of COVID-19.