Research map / 3D vision / Depth and stereo
Sky2Ground: A Benchmark for Site Modeling under Varying Altitude
Zengyan Wang, Sirshapan Mitra, Rajat Modi, Hui Lim, Yogesh Rawat
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 Masked-Satellite Attention (MSA) (ablation) | SkyNet · Sky2Ground G+S | avg RRA/RTA: 53.8 -> 62.7 | Improved |
| add Progressive-View Sampling (P-VS) (ablation) | SkyNet · Sky2Ground G+S | avg RRA/RTA: 53.8 -> 61.1 | Improved |
| add Curriculum-Aware Camera Sampling (CA-CS) (ablation) | SkyNet · Sky2Ground G+S | avg RRA/RTA: 53.8 -> 54.3 | Improved |
| add CA-CS on top of MSA (ablation) | SkyNet · Sky2Ground G+S | avg RRA/RTA: 62.7 -> 63.8 | Improved |
| add P-VS on top of MSA (ablation) | SkyNet · Sky2Ground G+S | avg RRA/RTA: 62.7 -> 64.9 | Improved |
| add CA-CS to MSA + P-VS (full model) (ablation) | SkyNet · Sky2Ground G+S | avg RRA/RTA: 64.9 -> 65.1 | Improved |
| also feed satellite images to the GAS encoder, not only Sat-Encoder (ablation) | SkyNet · Sky2Ground G+S | avg RRA/RTA: 53.1 -> 53.8 | Improved |
| replace additive satellite-token fusion with cross-attention (ablation) | SkyNet · Sky2Ground G+S | avg RRA/RTA: 65.1 -> 59.7 | Hurt |
| add 24 extra layers to VGGT (parameter-matched) (ablation) | VGGT fine-tuned · Sky2Ground G+S | avg RRA/RTA: 47.8 -> 50.4 | Mixed |
| fine-tune VGGT on Sky2Ground including satellite | VGGT · Sky2Ground GAS | avg RRA: 73.6 -> 55.4 | Hurt |
| fine-tune Dust3r on ground-aerial data | Dust3r · Sky2Ground GAS | avg RRA: 58.9 -> 64.1 | Improved |
| fine-tune Mast3r on ground-aerial data | Mast3r · Sky2Ground GAS | avg RRA: 53.2 -> 67.6 | 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.
- VGGT — VGGT: Visual Geometry Grounded Transformer (CVPR 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
- SkyScenes: A Synthetic Dataset for Aerial Scene Understanding (ECCV 2024 · similarity 0.86)
- Satellite to GroundScape - Large-scale Consistent Ground View Generation from Satellite Views (CVPR 2025 · similarity 0.86)
- AerialMegaDepth: Learning Aerial-Ground Reconstruction and View Synthesis (CVPR 2025 · similarity 0.86)
- Weather-Robust Cross-View Geo-Localization via Prototype-Based Semantic Part Discovery (NeurIPS 2026 · similarity 0.86)
- SkySense: A Multi-Modal Remote Sensing Foundation Model Towards Universal Interpretation for Earth Observation Imagery (CVPR 2024 · similarity 0.85)
- Horizon-GS: Unified 3D Gaussian Splatting for Large-Scale Aerial-to-Ground Scenes (CVPR 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.
