Research map / 3D vision / 3D generation
Sat2City: 3D City Generation from A Single Satellite Image with Cascaded Latent Diffusion
Tongyan Hua, Lutao Jiang, Ying-Cong Chen, Wufan Zhao
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 Re-Hash neck with dual-stage training instead of a unified latent (ablation) | Sat2City VAE · Sat2City dataset | color training loss | Improved |
| separate latent spaces for color via dual-stage training instead of unified geometry-color latent (ablation) | Sat2City VAE · Sat2City dataset | gradient conflict / color loss | Improved |
| add inverse sampling instead of direct color splatting (ablation) | Sat2City · Sat2City dataset | rendering artifacts (qualitative) | Improved |
| add dense latent diffusion stage to the cascade (ablation) | Sat2City · Sat2City dataset | generation quality (qualitative) | Improved |
| keep sparse latent stage supplying pruning decisions to appearance decoding (ablation) | Sat2City · Sat2City dataset | appearance decoding (qualitative) | Improved |
| use Sat2City instead of BlockFusion | Sat2City · Sat2City dataset | MMD (CD): 0.0326 -> 0.0165 | Improved |
| use Sat2City instead of BlockFusion | Sat2City · Sat2City dataset | COV (CD, %): 50.49 -> 100.00 | Improved |
| use Sat2City instead of BlockFusion | Sat2City · Sat2City dataset | COV (EMD, %): 55.66 -> 60.00 | Improved |
| direct 3D generation instead of Sat2Scene 2D rendering | Sat2City · user study | TPQ: 6.17 -> 7.35 | Improved |
| Sat2City instead of retrained Sat2Scene on low-density point clouds | Sat2City · Sat2City dataset (user study) | TSC: 3.30 -> 8.03 | Improved |
| direct 3D generation instead of CityDreamer 2D rendering | Sat2City · user study | TSC: 6.63 -> 8.03 | Improved |
Closest papers by meaning
- Sat2Scene: 3D Urban Scene Generation from Satellite Images with Diffusion (CVPR 2024 · similarity 0.87)
- CityDreamer: Compositional Generative Model of Unbounded 3D Cities (CVPR 2024 · similarity 0.87)
- Generative Gaussian Splatting for Unbounded 3D City Generation (CVPR 2025 · similarity 0.86)
- A 3D Scene is Worth 1K Tokens: 3D-Grounded Representation for Scene Generation at Scale (NeurIPS 2026 · similarity 0.86)
- LT3SD: Latent Trees for 3D Scene Diffusion (CVPR 2025 · similarity 0.86)
- CymbaDiff: Structured Spatial Diffusion for Sketch-based 3D Semantic Urban Scene Generation (NeurIPS 2025 · similarity 0.85)
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.
