AI Paper Atlas

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

SynAD: Enhancing Real-World End-to-End Autonomous Driving Models through Synthetic Data Integration

Jongsuk Kim, Jaeyoung Lee, Gyojin Han, Dong-Jae Lee, Minki Jeong, Junmo Kim

ICCV 2025 · Label-efficient and robust learning › Recognition and long-tail · approach: Dense features and foundation priors · 2 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
add synthetic scenes (100) to training (ablation)SynAD · nuScenes val planningplanning L2 avg: 1.15 -> 0.77Improved
add synthetic scenes (100) to training (ablation)SynAD · nuScenes val planningcollision rate avg: 0.25 -> 0.14Improved
scale synthetic scenes from 100 to 500 (same step) (ablation)SynAD · nuScenes val planningplanning L2 avg: 0.77 -> 0.85Hurt
scale synthetic scenes from 100 to 500 (same epoch) (ablation)SynAD · nuScenes val planningcollision rate avg: 0.13 -> 0.11Improved
scale synthetic scenes from 100 to 300 (same step) (ablation)SynAD · nuScenes val motion forecastingminFDE: 1.06 -> 1.02Improved
add synthetic map input x_SM to planning training (ablation)SynAD · nuScenes val planningplanning L2 avg: 1.15 -> 0.82Improved
add real map x_RM to planning on top of x_SM (ablation)SynAD · nuScenes val planningplanning L2 avg: 0.82 -> 0.79Improved
also update occupancy module with map input (ablation)SynAD · nuScenes val occupancyIoU-n: 60.2 -> 58.4Hurt
add proposed Map-to-BEV network instead of SwinUNETR (ablation)SynAD · nuScenes val planningplanning L2 avg: 1.08 -> 0.77Improved
add map and speed guides to agent guide (ablation)SynAD · nuScenes val planningcollision rate avg: 0.18 -> 0.14Improved
use longest ego selection instead of random (ablation)SynAD · nuScenes val planningcollision rate avg: 0.16 -> 0.14Improved
use SynAD with synthetic data versus baselines on planningSynAD · nuScenes val planningcollision rate avg: 0.71 -> 0.11Improved

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