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

In autonomous driving perception systems, 3D detection and tracking are thetwo fundamental tasks. This paper delves deeper into this field, building uponthe Sparse4D framework. We introduce two auxiliary training tasks (TemporalInstance Denoising and Quality Estimation) and propose decoupled attention tomake structural improvements, leading to significant enhancements in detectionperformance. Additionally, we extend the detector into a tracker using astraightforward approach that assigns instance ID during inference, furtherhighlighting the advantages of query-based algorithms. Extensive experimentsconducted on the nuScenes benchmark validate the effectiveness of the proposedimprovements. With ResNet50 as the backbone, we witnessed enhancements of3.0\%, 2.2\%, and 7.6\% in mAP, NDS, and AMOTA, achieving 46.9\%, 56.1\%, and49.0\%, respectively. Our best model achieved 71.9\% NDS and 67.7\% AMOTA onthe nuScenes test set. Code will be released at\url{https://github.com/linxuewu/Sparse4D}.

Sparse4D v3: Advancing End-to-End 3D Detection and Tracking