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

Aerial image analysis, specifically the semantic segmentation thereof, is thebasis for applications such as automatically creating and updating maps,tracking city growth, or tracking deforestation. In true orthophotos, which areoften used in these applications, many objects and regions can be approximatedwell by polygons. However, this fact is rarely exploited by state-of-the-artsemantic segmentation models. Instead, most models allow unnecessary degrees offreedom in their predictions by allowing arbitrary region shapes. We thereforepresent a refinement of our deep learning model which predicts binary spacepartitioning trees, an efficient polygon representation. The refinementsinclude a new feature decoder architecture and a new differentiable BSP treerenderer which both avoid vanishing gradients. Additionally, we designed anovel loss function specifically designed to improve the spatial partitioningdefined by the predicted trees. Furthermore, our expanded model can predictmultiple trees at once and thus can predict class-specific segmentations.Taking all modifications together, our model achieves state-of-the-artperformance while using up to 60% fewer model parameters when using a smallbackbone model or up to 20% fewer model parameters when using a large backbonemodel.

SegForestNet: Spatial-Partitioning-Based Aerial Image Segmentation