Research map / Video / Video generation
GeoV2V: Geometry-Grounded Video Diffusion Model for Driving Scene Generation
Yuchen Xi, tippy guo, Chenwei Hou, Zixu Liu, Jin Fang, Jason Liu, Ruigang Yang
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 |
|---|---|---|---|
| add source video conditioning (ablation) | GeoV2V · Para4D lane shifting 2 m | PSNR: 23.28 -> 23.88 | Improved |
| add source video conditioning (ablation) | GeoV2V · Para4D lane shifting 2 m | SSIM: 0.695 -> 0.697 | No clear effect |
| add dense depth geometry prior (ablation) | GeoV2V · Para4D lane shifting 2 m | PSNR: 17.18 -> 23.88 | Improved |
| add sparse LiDAR geometry prior (ablation) | GeoV2V · Para4D lane shifting 2 m | PSNR: 22.86 -> 23.88 | Improved |
| add both geometry priors (ablation) | GeoV2V · Para4D lane shifting 2 m | PSNR: 16.45 -> 23.88 | Improved |
| add dense depth prior (ablation) | GeoV2V · Para4D lane shifting 2 m | LPIPS: 0.124 -> 0.051 | Improved |
| train with synchronized multi-lane Para4D instead of misaligned side views (ablation) | GeoV2V · Para4D lane shifting 2 m | PSNR: 20.92 -> 23.88 | Improved |
| train with synchronized multi-lane Para4D (ablation) | GeoV2V · Para4D lane shifting 2 m | SSIM: 0.593 -> 0.697 | Improved |
| train with synchronized multi-lane Para4D (ablation) | GeoV2V · Para4D lane shifting 2 m | LPIPS: 0.102 -> 0.051 | Improved |
| GeoV2V vs EmerNeRF on dynamic interpolation | GeoV2V · Waymo dynamic interpolation | PSNR: 23.91 -> 23.42 | Hurt |
| GeoV2V vs StreetCrafter at 4 m real lane shift | GeoV2V · Real-world lane shifting 4 m | FID: 28.53 -> 32.58 | Hurt |
| GeoV2V vs StreetCrafter at 4 m real lane shift | GeoV2V · Real-world lane shifting 4 m | FVD: 369.64 -> 328.69 | Improved |
Closest papers by meaning
- Effective Multi-sensor Conditioning for Street-view Novel-view Synthesis (NeurIPS 2026 · similarity 0.90)
- DriveDreamer4D: World Models Are Effective Data Machines for 4D Driving Scene Representation (CVPR 2025 · similarity 0.90)
- RLGF: Reinforcement Learning with Geometric Feedback for Autonomous Driving Video Generation (NeurIPS 2025 · similarity 0.89)
- DriveScape: High-Resolution Driving Video Generation by Multi-View Feature Fusion (CVPR 2025 · similarity 0.89)
- Genesis: Multimodal Driving Scene Generation with Spatio-Temporal and Cross-Modal Consistency (NeurIPS 2025 · similarity 0.88)
- StreetCrafter: Street View Synthesis with Controllable Video Diffusion Models (CVPR 2025 · similarity 0.88)
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.
