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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Designing an efficient yet deployment-friendly 3D backbone to handle sparsepoint clouds is a fundamental problem in 3D perception. Compared with thecustomized sparse convolution, the attention mechanism in Transformers is moreappropriate for flexibly modeling long-range relationships and is easier to bedeployed in real-world applications. However, due to the sparse characteristicsof point clouds, it is non-trivial to apply a standard transformer on sparsepoints. In this paper, we present Dynamic Sparse Voxel Transformer (DSVT), asingle-stride window-based voxel Transformer backbone for outdoor 3Dperception. In order to efficiently process sparse points in parallel, wepropose Dynamic Sparse Window Attention, which partitions a series of localregions in each window according to its sparsity and then computes the featuresof all regions in a fully parallel manner. To allow the cross-set connection,we design a rotated set partitioning strategy that alternates between twopartitioning configurations in consecutive self-attention layers. To supporteffective downsampling and better encode geometric information, we also proposean attention-style 3D pooling module on sparse points, which is powerful anddeployment-friendly without utilizing any customized CUDA operations. Our modelachieves state-of-the-art performance with a broad range of 3D perceptiontasks. More importantly, DSVT can be easily deployed by TensorRT with real-timeinference speed (27Hz). Code will be available at\url{https://github.com/Haiyang-W/DSVT}.

DSVT: Dynamic Sparse Voxel Transformer with Rotated Sets