Research map / 3D vision / 3D generation
ScenDi: 3D-to-2D Scene Diffusion Cascades for Urban Generation
Hanlei Guo, Jiahao Shao, Xinya Chen, Xiyang Tan, Sheng Miao, Yujun Shen, Yiyi Liao
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 |
|---|---|---|---|
| condition 2D video diffusion on rendered 3DGS RGB instead of depth (ablation) | ScenDi (SVD) · KITTI-360 | FID: 78.6 -> 36.9 | Improved |
| condition 2D video diffusion on rendered 3DGS RGB instead of depth (ablation) | ScenDi (SVD) · Waymo | FID: 77.4 -> 41.3 | Improved |
| condition 2D video diffusion on rendered 3DGS RGB instead of depth (ablation) | ScenDi (SVD) · KITTI-360 | KID: 0.070 -> 0.026 | Improved |
| add BCE occupancy loss to VQ-VAE training (ablation) | Voxel-to-3DGS VQ-VAE · Waymo val | PSNR: 21.47 -> 21.56 | Improved |
| add BCE occupancy loss to VQ-VAE training (ablation) | Voxel-to-3DGS VQ-VAE · Waymo val | LPIPS: 0.327 -> 0.321 | Improved |
| increase Gaussians per voxel from 1 to 6 (ablation) | Voxel-to-3DGS VQ-VAE · Waymo val | PSNR: 21.56 -> 21.53 | No clear effect |
| increase Gaussians per voxel from 1 to 6 (ablation) | Voxel-to-3DGS VQ-VAE · Waymo val | LPIPS: 0.321 -> 0.317 | No clear effect |
| use WAN2.1-1.3B backbone instead of SVD for 2D refinement | ScenDi · KITTI-360 | FID: 36.9 -> 22.9 | Improved |
| add Diffusion Forcing to long-video inference (ablation) | ScenDi 2D video diffusion · clip transitions | background consistency (qualitative) | 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.
- DiscoScene — DisCoScene: Spatially Disentangled Generative Radiance Fields for Controllable 3D-Aware Scene Synthesis (CVPR 2023) · 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
- A 3D Scene is Worth 1K Tokens: 3D-Grounded Representation for Scene Generation at Scale (NeurIPS 2026 · similarity 0.89)
- LT3SD: Latent Trees for 3D Scene Diffusion (CVPR 2025 · similarity 0.88)
- Generative Gaussian Splatting for Unbounded 3D City Generation (CVPR 2025 · similarity 0.88)
- EVolSplat: Efficient Volume-based Gaussian Splatting for Urban View Synthesis (CVPR 2025 · similarity 0.87)
- SemCity: Semantic Scene Generation with Triplane Diffusion (CVPR 2024 · similarity 0.87)
- StarGen: A Spatiotemporal Autoregression Framework with Video Diffusion Model for Scalable and Controllable 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.
