Research map / 3D vision / Gaussian splatting
WildSplat: Feedforward Gaussian Splatting from Unposed In-the-Wild Images
Xiyu Zhang, Jingyu Zhuang, Hongjia Zhai, Zizheng Yan, Jinwei Chen, Guofeng Zhang, Qingnan Fan
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 WildSplat instead of AnySplat/WorldMirror (decoupled geometry and appearance) | feedforward GS · Phototourism (4 views) | PSNR: 15.61 -> 19.57 | Improved |
| use WildSplat instead of AnySplat | feedforward GS · MegaScenes (12 views) | PSNR: 16.31 -> 19.20 | Improved |
| add joint multi-reference supervision (ablation) | WildSplat · Phototourism (4 views) | PSNR: 16.88 -> 18.52 | Improved |
| add geometry-guided view sampling (ablation) | WildSplat · Phototourism (4 views) | PSNR: 17.99 -> 18.52 | Improved |
| add global pre-modulation (ablation) | WildSplat · Phototourism (12 views) | PSNR: 20.11 -> 20.15 | Improved |
| add global pre-modulation (ablation) | WildSplat · Phototourism (4 views) | PSNR: 18.25 -> 18.52 | Improved |
| replace AnySplat with WildSplat on consistent-lighting scenes | feedforward GS · DL3DV (6 views) | PSNR: 20.33 -> 22.37 | Improved |
| WildSplat pose head vs VGGT | WildSplat · MegaScenes pose | AUC@5: 0.310 vs 0.314 (VGGT) | No clear effect |
| WildSplat pose vs AnySplat | WildSplat · MegaScenes pose | AUC@5: 0.249 -> 0.310 | Improved |
| add appearance branch | WildSplat · 16-view inference cost | time / memory: 1.1 s -> 1.3 s; 8.2 GB -> 9.6 GB | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- VGGT — VGGT: Visual Geometry Grounded Transformer (CVPR 2025) · better on 0 of 1 shared result
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 0.91)
- MS-GS: Multi-Appearance Sparse-View 3D Gaussian Splatting in the Wild (NeurIPS 2025 · similarity 0.90)
- FreeSplat: Generalizable 3D Gaussian Splatting Towards Free View Synthesis of Indoor Scenes (NeurIPS 2024 · similarity 0.89)
- Robust Neural Rendering in the Wild with Asymmetric Dual 3D Gaussian Splatting (NeurIPS 2025 · similarity 0.89)
- WildGaussians: 3D Gaussian Splatting In the Wild (NeurIPS 2024 · similarity 0.89)
- WildCAT3D: Appearance-Aware Multi-View Diffusion in the Wild (NeurIPS 2025 · similarity 0.89)
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.
