English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Liver segmentation on images acquired using computed tomography (CT) andmagnetic resonance imaging (MRI) plays an important role in clinical managementof liver diseases. Compared to MRI, CT images of liver are more abundant andreadily available. However, MRI can provide richer quantitative information ofthe liver compared to CT. Thus, it is desirable to achieve unsupervised domainadaptation for transferring the learned knowledge from the source domaincontaining labeled CT images to the target domain containing unlabeled MRimages. In this work, we report a novel unsupervised domain adaptationframework for cross-modality liver segmentation via joint adversarial learningand self-learning. We propose joint semantic-aware and shape-entropy-awareadversarial learning with post-situ identification manner to implicitly alignthe distribution of task-related features extracted from the target domain withthose from the source domain. In proposed framework, a network is trained withthe above two adversarial losses in an unsupervised manner, and then a meancompleter of pseudo-label generation is employed to produce pseudo-labels totrain the next network (desired model). Additionally, semantic-awareadversarial learning and two self-learning methods, including pixel-adaptivemask refinement and student-to-partner learning, are proposed to train thedesired model. To improve the robustness of the desired model, a low-signalaugmentation function is proposed to transform MRI images as the input of thedesired model to handle hard samples. Using the public data sets, ourexperiments demonstrated the proposed unsupervised domain adaptation frameworkreached four supervised learning methods with a Dice score 0.912 plus or minus0.037 (mean plus or minus standard deviation).

Unsupervised domain adaptation for cross-modality liver segmentation via joint adversarial learning and self-learning