Research map / 3D vision / Gaussian splatting
Gaussian in the wild: 3D Gaussian Splatting for Unconstrained Image Collections
Dongbin Zhang, Chuming Wang, Weitao Wang, Peihao Li, Minghan Qin, Haoqian Wang
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 per-point adaptive sampling of K feature maps instead of fixed coordinates (ablation) | GS-W · PhotoTourism Sacre Coeur | PSNR: 21.45 -> 23.24 | Improved |
| add K feature maps (ablation) | GS-W · PhotoTourism Brandenburg Gate | PSNR: 26.37 -> 27.96 | Improved |
| add projection feature map (ablation) | GS-W · PhotoTourism Brandenburg Gate | PSNR: 26.61 -> 27.96 | Improved |
| add visibility map for transient objects (ablation) | GS-W · PhotoTourism Brandenburg Gate | PSNR: 28.64 -> 27.96 | Hurt |
| separate intrinsic and dynamic appearance features (ablation) | GS-W · PhotoTourism Sacre Coeur | LPIPS: 0.1469 -> 0.1300 | Improved |
| separate intrinsic and dynamic appearance features (ablation) | GS-W · PhotoTourism Brandenburg Gate | PSNR: 28.22 -> 27.96 | Mixed |
| use per-point dynamic features instead of a single global image feature (ablation) | GS-W · PhotoTourism average of 3 scenes | PSNR: 22.43 -> 24.70 | Improved |
| increase number of feature maps K beyond 3 (ablation) | GS-W · PhotoTourism Sacre Coeur | PSNR: 23.24 -> 22.94 (K=8) | No clear effect |
| cache appearance features per Gaussian at inference | GS-W · PhotoTourism Brandenburg Gate | FPS: 55.8 -> 221 | Improved |
| use 3D Gaussians with appearance features instead of NeRF-based in-the-wild methods | CR-NeRF · PhotoTourism Brandenburg Gate | FPS: 0.0445 -> 55.8 | Improved |
Closest papers by meaning
- Wild-GS: Real-Time Novel View Synthesis from Unconstrained Photo Collections (NeurIPS 2024 · similarity 0.94)
- MS-GS: Multi-Appearance Sparse-View 3D Gaussian Splatting in the Wild (NeurIPS 2025 · similarity 0.91)
- COLMAP-Free 3D Gaussian Splatting (CVPR 2024 · similarity 0.90)
- Structure Consistent Gaussian Splatting with Matching Prior for Few-shot Novel View Synthesis (NeurIPS 2024 · similarity 0.90)
- WildGaussians: 3D Gaussian Splatting In the Wild (NeurIPS 2024 · similarity 0.90)
- Textured Gaussians for Enhanced 3D Scene Appearance Modeling (CVPR 2025 · similarity 0.90)
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.
