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

3D object detectors usually rely on hand-crafted proxies, e.g., anchors orcenters, and translate well-studied 2D frameworks to 3D. Thus, sparse voxelfeatures need to be densified and processed by dense prediction heads, whichinevitably costs extra computation. In this paper, we instead propose VoxelNextfor fully sparse 3D object detection. Our core insight is to predict objectsdirectly based on sparse voxel features, without relying on hand-craftedproxies. Our strong sparse convolutional network VoxelNeXt detects and tracks3D objects through voxel features entirely. It is an elegant and efficientframework, with no need for sparse-to-dense conversion or NMS post-processing.Our method achieves a better speed-accuracy trade-off than other mainframedetectors on the nuScenes dataset. For the first time, we show that a fullysparse voxel-based representation works decently for LIDAR 3D object detectionand tracking. Extensive experiments on nuScenes, Waymo, and Argoverse2benchmarks validate the effectiveness of our approach. Without bells andwhistles, our model outperforms all existing LIDAR methods on the nuScenestracking test benchmark.

VoxelNeXt: Fully Sparse VoxelNet for 3D Object Detection and Tracking