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

Although domain adaptation has been extensively studied in naturalimage-based segmentation task, the research on cross-domain segmentation forvery high resolution (VHR) remote sensing images (RSIs) still remainsunderexplored. The VHR RSIs-based cross-domain segmentation mainly faces twocritical challenges: 1) Large area land covers with many diverse objectcategories bring severe local patch-level data distribution deviations, thusyielding different adaptation difficulties for different local patches; 2)Different VHR sensor types or dynamically changing modes cause the VHR imagesto go through intensive data distribution differences even for the samegeographical location, resulting in different global feature-level domain gap.To address these challenges, we propose a curriculum-style local-to-globalcross-domain adaptation framework for the segmentation of VHR RSIs. Theproposed curriculum-style adaptation performs the adaptation process in aneasy-to-hard way according to the adaptation difficulties that can be obtainedusing an entropy-based score for each patch of the target domain, and thus wellaligns the local patches in a domain image. The proposed local-to-globaladaptation performs the feature alignment process from the locally semantic toglobally structural feature discrepancies, and consists of a semantic-leveldomain classifier and an entropy-level domain classifier that can reduce theabove cross-domain feature discrepancies. Extensive experiments have beenconducted in various cross-domain scenarios, including geographic locationvariations and imaging mode variations, and the experimental resultsdemonstrate that the proposed method can significantly boost the domainadaptability of segmentation networks for VHR RSIs. Our code is available at:https://github.com/BOBrown/CCDA_LGFA.

Curriculum-style Local-to-global Adaptation for Cross-domain Remote Sensing Image Segmentation