Research map / 3D vision / Gaussian splatting
Surface Reconstruction for 3D Gaussian Splatting via Local Structural Hints
Qianyi Wu, Jianmin Zheng, Jianfei Cai
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 monocular normal guidance to Scaffold-GS (ablation) | Scaffold-GS · Replica | F-score: 55.33 -> 65.28 (depth-only vs normal-only) | Improved |
| add monocular depth guidance on top of normal guidance (ablation) | Scaffold-GS+N · Replica | F-score: 65.28 -> 64.00 | Hurt |
| add joint neural implicit RIMLS regularization (ablation) | Scaffold-GS+N · Replica | Normal consistency: 80.03 -> 83.99 | Improved |
| use RIMLS instead of IMLS in the joint regularizer (ablation) | GSrec · Replica | F-score: 65.46 -> 66.04 | Improved |
| add depth guidance with joint RIMLS (full model) (ablation) | GSrec · Replica | F-score: 66.04 -> 67.22 | Improved |
| add gradient term to joint loss (ablation) | GSrec · Replica | F-score: 63.56 -> 67.22 | Improved |
| add MLS term to joint loss (ablation) | GSrec · Replica | F-score: 64.55 -> 67.22 | Improved |
| add eikonal term to joint loss (ablation) | GSrec · Replica | F-score: 66.62 -> 67.22 | Improved |
| increase number of Gaussians L used in RIMLS (ablation) | GSrec · Replica | F-score: 64.52 (L=1) -> 67.22 (L=50) | Improved |
| replace SuGaR regularization with GSrec on 3DGS surface extraction | GSrec vs SuGaR (SDF) · Replica | Chamfer distance: 9.57 -> 7.08 | Improved |
| GSrec vs neural implicit surface methods on ScanNet | GSrec · ScanNet | F-score: 0.599 vs MonoSDF (MLP) 0.733; training under 1 hour vs 12+ hours | Mixed |
Closest papers by meaning
- SuGaR: Surface-Aligned Gaussian Splatting for Efficient 3D Mesh Reconstruction and High-Quality Mesh Rendering (CVPR 2024 · similarity 0.93)
- VA-GS: Enhancing the Geometric Representation of Gaussian Splatting via View Alignment (NeurIPS 2025 · similarity 0.92)
- FatesGS: Fast and Accurate Sparse-View Surface Reconstruction Using Gaussian Splatting with Depth-Feature Consistency (AAAI 2025 · similarity 0.92)
- Sparse2DGS: Geometry-Prioritized Gaussian Splatting for Surface Reconstruction from Sparse Views (CVPR 2025 · similarity 0.91)
- Neural Signed Distance Function Inference through Splatting 3D Gaussians Pulled on Zero-Level Set (NeurIPS 2024 · similarity 0.91)
- Structure Consistent Gaussian Splatting with Matching Prior for Few-shot Novel View Synthesis (NeurIPS 2024 · similarity 0.91)
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.
