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

This paper aims for high-performance offline LiDAR-based 3D object detection.We first observe that experienced human annotators annotate objects from atrack-centric perspective. They first label the objects with clear shapes in atrack, and then leverage the temporal coherence to infer the annotations ofobscure objects. Drawing inspiration from this, we propose a high-performanceoffline detector in a track-centric perspective instead of the conventionalobject-centric perspective. Our method features a bidirectional tracking moduleand a track-centric learning module. Such a design allows our detector to inferand refine a complete track once the object is detected at a certain moment. Werefer to this characteristic as "onCe detecTed, neveR Lost" and name theproposed system CTRL. Extensive experiments demonstrate the remarkableperformance of our method, surpassing the human-level annotating accuracy andthe previous state-of-the-art methods in the highly competitive Waymo OpenDataset without model ensemble. The code will be made publicly available athttps://github.com/tusen-ai/SST.

Once Detected, Never Lost: Surpassing Human Performance in Offline LiDAR based 3D Object Detection