Research map / 3D vision / 3D generation
SatDreamer360: Multiview-Consistent Generation of Ground-Level Scenes from Satellite Imagery
Xianghui Ze, Beiyi Zhu, Zhenbo Song, Jianfeng Lu, Yujiao Shi
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 |
|---|---|---|---|
| use triplane representation instead of BEV (XY-plane only) (ablation) | SatDreamer360 · VIGOR | PSNR: 13.64 -> 14.08 | Improved |
| use triplane instead of BEV (ablation) | SatDreamer360 · VIGOR | Depth error: 7.061 -> 6.727 | Improved |
| use ray-based pixel attention instead of standard cross-attention (ablation) | SatDreamer360 · CVUSA | LPIPS (Alex): 0.5413 -> 0.3469 | Improved |
| use ray-based pixel attention instead of standard cross-attention (ablation) | SatDreamer360 · CVUSA | time (s): 120.28 -> 39.64 | Improved |
| use ray-based pixel attention instead of standard cross-attention (ablation) | SatDreamer360 · CVUSA | memory (MB): 22506 -> 18296 | Improved |
| add dynamic refinement of ray sampling offsets (ablation) | Ray-Based Pixel Attention · CVUSA | PSNR: 13.45 -> 13.88 | Improved |
| add dynamic refinement of ray sampling offsets (ablation) | Ray-Based Pixel Attention · CVUSA | Depth error: 19.69 -> 19.36 | Improved |
| use sparse (two preceding frames) instead of dense interframe attention (ablation) | SatDreamer360 · VIGOR++ | FVD: 2.253 -> 2.101 | Improved |
| use sparse instead of dense interframe attention (ablation) | SatDreamer360 · VIGOR++ | PSNR: 12.91 -> 12.75 | No clear effect |
| use sparse instead of dense interframe attention (ablation) | SatDreamer360 · VIGOR++ (30 frames) | time (s): 120.75 -> 32.71 | Improved |
Closest papers by meaning
- Satellite to GroundScape - Large-scale Consistent Ground View Generation from Satellite Views (CVPR 2025 · similarity 0.88)
- AerialMegaDepth: Learning Aerial-Ground Reconstruction and View Synthesis (CVPR 2025 · similarity 0.85)
- SGFormer: Satellite-Ground Fusion for 3D Semantic Scene Completion (CVPR 2025 · similarity 0.85)
- BevSplat: Resolving Height Ambiguity via Feature-Based Gaussian Primitives for Weakly-Supervised Cross-View Localization (NeurIPS 2025 · similarity 0.85)
- Seeing Across Skies and Streets: Feedforward 3D Reconstruction from Satellite, Drone, and Ground Images (NeurIPS 2026 · similarity 0.84)
- Controllable Satellite-to-Street-View Synthesis with Precise Pose Alignment and Zero-Shot Environmental Control (ICLR 2025 · similarity 0.84)
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.
