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

Automatic and periodic recompiling of building databases with up-to-datehigh-resolution images has become a critical requirement for rapidly developingurban environments. However, the architecture of most existing approaches forchange extraction attempts to learn features related to changes but ignoresobjectives related to buildings. This inevitably leads to the generation ofsignificant pseudo-changes, due to factors such as seasonal changes in imagesand the inclination of building fa\c{c}ades. To alleviate the above-mentionedproblems, we developed a contrastive learning approach by validating historicalbuilding footprints against single up-to-date remotely sensed images. Thiscontrastive learning strategy allowed us to inject the semantics of buildingsinto a pipeline for the detection of changes, which is achieved by increasingthe distinguishability of features of buildings from those of non-buildings. Inaddition, to reduce the effects of inconsistencies between historical buildingpolygons and buildings in up-to-date images, we employed a deformableconvolutional neural network to learn offsets intuitively. In summary, weformulated a multi-branch building extraction method that identifies newlyconstructed and removed buildings, respectively. To validate our method, weconducted comparative experiments using the public Wuhan University buildingchange detection dataset and a more practical dataset named SI-BU that weestablished. Our method achieved F1 scores of 93.99% and 70.74% on the abovedatasets, respectively. Moreover, when the data of the public dataset weredivided in the same manner as in previous related studies, our method achievedan F1 score of 94.63%, which surpasses that of the state-of-the-art method.

BCE-Net: Reliable Building Footprints Change Extraction based on Historical Map and Up-to-Date Images using Contrastive Learning