Research map / 3D vision / 3D generation
Sat3DGen: Comprehensive Street-Level 3D Scene Generation from Single Satellite Image
Ming Qian, Zimin Xia, Changkun Liu, Shuailei Ma, Wen Wang, Zeran Ke, Bin Tan, Hang Zhang, Gui-Song Xia
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 L_dep, L_grav and spatial tokens to canonical image-to-3D baseline (ablation) | Canonical Image-to-3D · VIGOR-OOD | FID: 35.6 -> 21.6 | Improved |
| add gravity-based density variation loss (ablation) | Sat3DGen base · VIGOR-OOD | FID: 25.9 -> 21.6 | Improved |
| add gravity-based density variation loss (ablation) | Sat3DGen base · VIGOR-OOD | RMSE (m): 5.21 -> 5.23 | No clear effect |
| add monocular relative-depth prior in satellite view (ablation) | Sat3DGen base · VIGOR-OOD | RMSE (m): 5.75 -> 5.23 | Improved |
| add monocular relative-depth prior in satellite view (ablation) | Sat3DGen base · VIGOR-OOD | FID: 23.7 -> 21.6 | Improved |
| add spatial tokens padding (ablation) | Sat3DGen base · VIGOR-OOD | RMSE (m): 5.64 -> 5.23 | Improved |
| add spatial tokens padding (ablation) | Sat3DGen base · VIGOR-OOD | FID: 24.8 -> 21.6 | Improved |
| add perspective-view training from panoramas (ablation) | Sat3DGen base · VIGOR-OOD | FID: 21.6 -> 19.2 | Improved |
| add perspective-view training from panoramas (ablation) | Sat3DGen base · VIGOR-OOD | RMSE (m): 5.23 -> 5.20 | No clear effect |
| geometry-first design vs Sat2Density++ | Sat3DGen · VIGOR-OOD DSM | RMSE (m): 6.76 -> 5.20 | Improved |
| geometry-first design vs ControlS2S under OOD city split | Sat3DGen · VIGOR-OOD | PSNR / SSIM: 13.80 -> 12.83 PSNR | Hurt |
Closest papers by meaning
- Sat2City: 3D City Generation from A Single Satellite Image with Cascaded Latent Diffusion (ICCV 2025 · similarity 0.89)
- Geometry-Guided Street-View Panorama Synthesis From Satellite Imagery (TPAMI 2022 · similarity 0.88)
- SS3DM: Benchmarking Street-View Surface Reconstruction with a Synthetic 3D Mesh Dataset (NeurIPS 2024 · similarity 0.87)
- CymbaDiff: Structured Spatial Diffusion for Sketch-based 3D Semantic Urban Scene Generation (NeurIPS 2025 · similarity 0.86)
- GeoLRM: Geometry-Aware Large Reconstruction Model for High-Quality 3D Gaussian Generation (NeurIPS 2024 · similarity 0.85)
- AerialMegaDepth: Learning Aerial-Ground Reconstruction and View Synthesis (CVPR 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.
