Research map / 3D vision / Gaussian splatting
FreeGen: Feed-Forward Reconstruction–Generation Co-Training for Free-Viewpoint Driving Scene Synthesis
Shijie Chen, Peixi Peng
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 |
|---|---|---|---|
| add generative refinement to reconstruction (ablation) | FreeGen · nuScenes left 2m shift | FVD: 246.73 -> 60.53 | Improved |
| add generative refinement to reconstruction (ablation) | FreeGen · nuScenes left 2m shift | FID: 23.25 -> 15.49 | Improved |
| add opacity map as geometry condition (ablation) | FreeGen · nuScenes left 2m shift | FID: 15.49 -> 13.86 | Improved |
| add opacity map as geometry condition (ablation) | FreeGen · nuScenes left 2m shift | FVD: 60.53 -> 57.21 | Improved |
| add reconstruction co-training (ablation) | FreeGen · nuScenes left 2m shift | FID: 13.86 -> 14.39 | Hurt |
| add reconstruction co-training (ablation) | FreeGen · nuScenes left 2m shift | FVD: 57.21 -> 54.93 | Improved |
| add generation co-training (ablation) | FreeGen · nuScenes left 2m shift | FID: 13.86 -> 13.16 | Improved |
| add generation co-training (ablation) | FreeGen · nuScenes left 2m shift | FVD: 57.21 -> 51.78 | Improved |
| co-train both reconstruction and generation (ablation) | FreeGen · nuScenes left 2m shift | FVD: 57.21 -> 50.92 | Improved |
| use our feed-forward reconstruction geometry condition instead of image warping (ablation) | FreeGen · nuScenes shift ±2m | FID: 60.64 -> 20.96 | Improved |
| use our feed-forward reconstruction geometry condition instead of image warping (ablation) | FreeGen · nuScenes shift ±4m | FVD: 398.36 -> 370.79 | Improved |
| use reconstruction model geometry instead of image warping on recorded trajectory (ablation) | FreeGen · nuScenes original trajectory | PSNR: 18.14 -> 20.40 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- OmniRe — OmniRe: Omni Urban Scene Reconstruction (ICLR 2025) · better on 6 of 6 shared results
- FreeVS — FreeVS: Generative View Synthesis on Free Driving Trajectory (ICLR 2025) · better on 2 of 2 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
- FreeVS: Generative View Synthesis on Free Driving Trajectory (ICLR 2025 · similarity 0.92)
- FreeSim: Toward Free-viewpoint Camera Simulation in Driving Scenes (CVPR 2025 · similarity 0.92)
- VR-Drive: Viewpoint-Robust End-to-End Driving with Feed-Forward 3D Gaussian Splatting (NeurIPS 2025 · similarity 0.89)
- X-Scene: Large-Scale Driving Scene Generation with High Fidelity and Flexible Controllability (NeurIPS 2025 · similarity 0.88)
- FRUC: Feedforward Dynamic Scene Reconstruction from Uncalibrated Collaborative Driving Views (NeurIPS 2026 · similarity 0.87)
- UniScene: Unified Occupancy-centric Driving Scene Generation (CVPR 2025 · 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.
