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

Unsupervised domain adaptation has attracted growing research attention onsemantic segmentation. However, 1) most existing models cannot be directlyapplied into lesions transfer of medical images, due to the diverse appearancesof same lesion among different datasets; 2) equal attention has been paid intoall semantic representations instead of neglecting irrelevant knowledge, whichleads to negative transfer of untransferable knowledge. To address thesechallenges, we develop a new unsupervised semantic transfer model including twocomplementary modules (i.e., T_D and T_F ) for endoscopic lesions segmentation,which can alternatively determine where and how to explore transferabledomain-invariant knowledge between labeled source lesions dataset (e.g.,gastroscope) and unlabeled target diseases dataset (e.g., enteroscopy).Specifically, T_D focuses on where to translate transferable visual informationof medical lesions via residual transferability-aware bottleneck, whileneglecting untransferable visual characterizations. Furthermore, T_F highlightshow to augment transferable semantic features of various lesions andautomatically ignore untransferable representations, which exploresdomain-invariant knowledge and in return improves the performance of T_D. Tothe end, theoretical analysis and extensive experiments on medical endoscopicdataset and several non-medical public datasets well demonstrate thesuperiority of our proposed model.

What Can Be Transferred: Unsupervised Domain Adaptation for Endoscopic Lesions Segmentation