Research map / 3D vision / 3D generation
ReinDriveGen: Reinforcement Post-Training for Out-of-Distribution Driving Scene Generation
Hao ZHANG, Lue Fan, Weikang Bian, Zehuan Wu, Lewei Lu, Zhaoxiang Zhang, Hongsheng LI
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 changed | Where | Measured | Effect |
|---|---|---|---|
| use pairwise preference model instead of pointwise reward model (ablation) | vehicle preference model · Preference model test set | preference accuracy (%): 91 -> 95 | Improved |
| use pairwise win-rate reward instead of pointwise reward in RL post-training (ablation) | ReinDriveGen RL post-training · 5 OOD scenarios (qualitative) | vehicle quality (visual) | Improved |
| add vehicle completion (AdaPoinTr) to pseudo-images (ablation) | ReinDriveGen · Vehicle trajectory editing (qualitative) | vehicle geometry and texture (visual) | Improved |
| add RL post-training (DiffusionNFT with pairwise reward) after SFT (ablation) | ReinDriveGen · OOD edit conditions | evaluation score P(RL beats SFT): 0.5 at step 0, rising above 0.5 | Improved |
| ReinDriveGen vs StreetCrafter on trajectory editing | ReinDriveGen · Vehicle trajectory editing (5 m shift) | Motion smoothness: 0.981715 -> 0.977355 | Hurt |
| ReinDriveGen vs best baseline on novel ego trajectory | ReinDriveGen · Waymo lane-change 4 m shift | FID: 66.93 -> 51.99 | Improved |
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.
- Street Gaussians — Street Gaussians: Modeling Dynamic Urban Scenes with Gaussian Splatting (ECCV 2024) · better on 1 of 1 shared result
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
- RLGF: Reinforcement Learning with Geometric Feedback for Autonomous Driving Video Generation (NeurIPS 2025 · similarity 0.88)
- DriveEditor: A Unified 3D Information-Guided Framework for Controllable Object Editing in Driving Scenes (AAAI 2025 · similarity 0.88)
- SceneCrafter: Controllable Multi-View Driving Scene Editing (CVPR 2025 · similarity 0.88)
- X-Scene: Large-Scale Driving Scene Generation with High Fidelity and Flexible Controllability (NeurIPS 2025 · similarity 0.87)
- Generating Multimodal Driving Scenes via Next-Scene Prediction (CVPR 2025 · similarity 0.87)
- SimGen: Simulator-conditioned Driving Scene Generation (NeurIPS 2024 · similarity 0.87)
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.
