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

Semantic segmentation with Convolutional Neural Networks is amemory-intensive task due to the high spatial resolution of feature maps andoutput predictions. In this paper, we present Quadtree Generating Networks(QGNs), a novel approach able to drastically reduce the memory footprint ofmodern semantic segmentation networks. The key idea is to use quadtrees torepresent the predictions and target segmentation masks instead of dense pixelgrids. Our quadtree representation enables hierarchical processing of an inputimage, with the most computationally demanding layers only being used atregions in the image containing boundaries between classes. In addition, givena trained model, our representation enables flexible inference schemes totrade-off accuracy and computational cost, allowing the network to adapt inconstrained situations such as embedded devices. We demonstrate the benefits ofour approach on the Cityscapes, SUN-RGBD and ADE20k datasets. On Cityscapes, weobtain an relative 3% mIoU improvement compared to a dilated network withsimilar memory consumption; and only receive a 3% relative mIoU drop comparedto a large dilated network, while reducing memory consumption by over4$\times$.

Quadtree Generating Networks: Efficient Hierarchical Scene Parsing with Sparse Convolutions