Domain adaptation based self-correction model for COVID-19 infection segmentation in CT images. (15th August 2021)

• Record Type: Journal Article
• Authors: Jin, Qiangguo; Cui, Hui; Sun, Changming; Meng, Zhaopeng; Wei, Leyi; Su, Ran
• Is Part Of: Expert systems with applications. Volume 176(2021)
• Publication Date: 2021-08-15
• Subjects: COVID-19 CT segmentation -- Domain adaptation -- Self-correction learning -- Attention mechanism; Expert systems (Computer science) -- Periodicals; Systèmes experts (Informatique) -- Périodiques; Electronic journals; 006.33
• DOI: 10.1016/j.eswa.2021.114848
• Languages: English
• ISSNs: 0957-4174

Abstract

Highlights: Address the domain shift problem with a limited amount of available datasets. An attention and feature domain enhanced model to segment COVID infection. A dual-domain enhanced self-correction learning algorithm to refine results. Comprehensive evaluation on three cross-site COVID datasets. Abstract: The capability of generalization to unseen domains is crucial for deep learning models when considering real-world scenarios. However, current available medical image datasets, such as those for COVID-19 CT images, have large variations of infections and domain shift problems. To address this issue, we propose a prior knowledge driven domain adaptation and a dual-domain enhanced self-correction learning scheme. Based on the novel learning scheme, a domain adaptation based self-correction model (DASC-Net) is proposed for COVID-19 infection segmentation on CT images. DASC-Net consists of a novel attention and feature domain enhanced domain adaptation model (AFD-DA) to solve the domain shifts and a self-correction learning process to refine segmentation results. The innovations in AFD-DA include an image-level activation feature extractor with attention to lung abnormalities and a multi-level discrimination module for hierarchical feature domain alignment. The proposed self-correction learning process adaptively aggregates the learned model and corresponding pseudo labels for the propagation of aligned source and target domain information to alleviate the overfitting to noises caused by pseudo labels. Extensive experiments over three publicly available COVID-19 CT datasets demonstrate that DASC-Net consistently outperforms state-of-the-art segmentation, domain shift, and coronavirus infection segmentation methods. Ablation analysis further shows the effectiveness of the major components in our model. The DASC-Net enriches the theory of domain adaptation and self-correction learning in medical imaging and can be generalized to multi-site COVID-19 infection segmentation on CT images for clinical deployment.