Research map / 3D vision / Gaussian splatting
Scene Grounding in the Wild
Tamir Cohen, Leo Segre, Shay Shomer-Chai, Shai Avidan, Hadar Averbuch-Elor
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 semantic inverse optimization on COLMAP init | COLMAP · WikiEarth | rotation error (deg): 4.99 -> 2.48 | Improved |
| add semantic inverse optimization on gDLS+++ init | gDLS+++ · WikiEarth | rotation error (deg): 2.86 -> 2.69 (translation 0.12 -> 0.13) | Mixed |
| add semantic inverse optimization on SP+LG init | SP+LG · WikiEarth | rotation error (deg): 3.74 -> 3.13 | Improved |
| use semantic DINOv2 features instead of photometric loss (ablation) | Ours (COLMAP init) · WikiEarth | rotation error (deg): 6.48 -> 2.48 | Improved |
| use DINOv2 instead of LSeg features (ablation) | Ours (COLMAP init) · WikiEarth | translation error: 0.34 -> 0.12 | Improved |
| use DINOv2 instead of DINOv2 + DVT (ablation) | Ours (COLMAP init) · WikiEarth | rotation error (deg): 2.86 -> 2.48 | Improved |
| add Least Trimmed Squares robust optimization (ablation) | Ours (COLMAP init) · WikiEarth | rotation error (deg): 3.78 -> 2.48 | Improved |
| soft (per-iteration) LTS instead of fixed LTS (ablation) | Ours (COLMAP init) · WikiEarth | rotation error (deg): 2.78 -> 2.48 | Improved |
| LTS instead of IRLS (ablation) | Ours (COLMAP init) · WikiEarth | rotation error (deg): 3.51 -> 2.48 | Improved |
| L1 feature loss instead of L2 (ablation) | Ours (COLMAP init) · WikiEarth | rotation error (deg): 4.21 -> 2.48 | Improved |
| ground to reference model instead of feed-forward 3D models | VGGT/MASt3R/DUSt3R/pi3 · WikiEarth meta-to-reference | rotation error (deg): 24.18 (best baseline MASt3R) -> 2.59 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 on most shared results.
- gDLS+++ — gDLS*: Generalized Pose-and-Scale Estimation Given Scale and Gravity Priors (CVPR 2020) · better on 3 of 4 shared results
- VGGT — VGGT: Visual Geometry Grounded Transformer (CVPR 2025) · better on 2 of 2 shared results
- DUSt3R — DUSt3R: Geometric 3D Vision Made Easy (CVPR 2024) · 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
- Wild-GS: Real-Time Novel View Synthesis from Unconstrained Photo Collections (NeurIPS 2024 · similarity 0.88)
- PF3plat: Pose-Free Feed-Forward 3D Gaussian Splatting for Novel View Synthesis (ICML 2025 · similarity 0.88)
- MS-GS: Multi-Appearance Sparse-View 3D Gaussian Splatting in the Wild (NeurIPS 2025 · similarity 0.87)
- No Pose, No Problem: Surprisingly Simple 3D Gaussian Splats from Sparse Unposed Images (ICLR 2025 · similarity 0.87)
- Horizon-GS: Unified 3D Gaussian Splatting for Large-Scale Aerial-to-Ground Scenes (CVPR 2025 · similarity 0.87)
- AerialMegaDepth: Learning Aerial-Ground Reconstruction and View Synthesis (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.
