AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

VAD: Vectorized Scene Representation for Efficient Autonomous Driving

Bo Jiang, Shaoyu Chen, Qing Xu, Bencheng Liao, Jiajie Chen, Helong Zhou, Qian Zhang, Wenyu Liu, Chang Huang, Xinggang Wang

ICCV 2023 · Label-efficient and robust learning › Recognition and long-tail · approach: Dense features and foundation priors · 478 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use vectorized scene representation and planning instead of UniAD's rasterized pipelineVAD-Base · nuScenes val open-loop planningaverage collision rate (%): 0.31 -> 0.22Improved
use vectorized scene representation and planning instead of UniAD's rasterized pipelineVAD-Base · nuScenes val open-loop planningaverage L2 error (m): 1.03 -> 0.72Improved
drop auxiliary tasks (tracking, occupancy) and post-processing with vectorized representation (VAD-Tiny)VAD · nuScenes val open-loop planningFPS: 1.8 -> 16.8Improved
use vectorized map instead of rasterized map representation (no map constraints) (ablation)VAD · nuScenes valaverage collision rate (%): 0.39 -> 0.26Improved
use vectorized map instead of rasterized map representation (no map constraints) (ablation)VAD · nuScenes valaverage L2 error (m): 0.74 -> 0.77Hurt
add ego-map interaction in planning (ablation)VAD · nuScenes valaverage L2 error (m): 0.86 -> 0.72Improved
add ego-agent interaction in planning (ablation)VAD · nuScenes valaverage collision rate (%): 0.26 -> 0.22Improved
add all three vectorized planning constraints (ablation)VAD · nuScenes valaverage collision rate (%): 0.28 -> 0.22Improved
add ego-boundary overstepping constraint only (ablation)VAD · nuScenes valaverage collision rate (%): 0.28 -> 0.24Improved
add ego-agent collision constraint only (ablation)VAD · nuScenes valaverage L2 error (m): 0.76 -> 0.77No clear effect
give ego status as input to planningVAD-Base · nuScenes val open-loop planningaverage L2 error (m): 0.72 -> 0.37Improved
use VAD-Base instead of UniAD on 1s collision rateVAD-Base · nuScenes val open-loop planning1s collision rate (%): 0.05 (UniAD) vs 0.07 (VAD-Base)Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy