Research map / 3D vision / Gaussian splatting
Incremental Online Scene Reconstruction by 3D Gaussian Triangulation
Yanjin Zhu, Shaofan Liu, Jianke Zhu
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 planar constraint (ablation) | full method · Replica Office1 ablation | Accuracy (cm): 1.19 -> 0.88 | Improved |
| add neighborhood selection in triangulation (ablation) | full method · Replica Office1 ablation | Accuracy (cm): 1.26 -> 0.88 | Improved |
| add direct triangulation instead of Poisson reconstruction (ablation) | full method · Replica Office1 ablation | Accuracy (cm): 0.89 -> 0.88 | Improved |
| add local remeshing (ablation) | full method · Replica Office1 ablation | Accuracy (cm): 1.32 -> 0.88 | Mixed |
| add region freezing (ablation) | full method · Replica Office1 ablation | Accuracy (cm): 0.92 -> 0.88 | Improved |
| add sparse loss to the baseline losses (ablation) | baseline losses · Replica Office1 ablation | Accuracy (cm): 1.13 -> 1.09 | Improved |
| add plane loss and normal loss on top of sparse loss (ablation) | full losses · Replica Office1 ablation | Accuracy (cm): 1.09 -> 0.88 | Improved |
| use full losses instead of baseline (ablation) | baseline losses · Replica Office1 ablation | Completion ratio: 86.34 -> 88.32 | Improved |
| increase tau in geometric Gaussian selection (ablation) | geometric Gaussian set · Replica | Completion | Mixed |
| use tau 0.0001 (too small) (ablation) | geometric Gaussian set · Replica | Mesh completeness | Hurt |
| add incremental freezing of optimized regions vs baselines (ablation) | Ours · Replica Office0 | Peak memory (MB) and FPS: 2325 MB, 10.34 FPS | 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.
- NICE-SLAM — NICE-SLAM: Neural Implicit Scalable Encoding for SLAM (CVPR 2022) · 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
- SuGaR: Surface-Aligned Gaussian Splatting for Efficient 3D Mesh Reconstruction and High-Quality Mesh Rendering (CVPR 2024 · similarity 0.91)
- Evolving High-Quality Rendering and Reconstruction in a Unified Framework with Contribution-Adaptive Regularization (CVPR 2025 · similarity 0.90)
- Dynamic Gaussians Mesh: Consistent Mesh Reconstruction from Dynamic Scenes (ICLR 2025 · similarity 0.89)
- GaussianPro: 3D Gaussian Splatting with Progressive Propagation (ICML 2024 · similarity 0.89)
- GaussianSpa: An "Optimizing-Sparsifying" Simplification Framework for Compact and High-Quality 3D Gaussian Splatting (CVPR 2025 · similarity 0.89)
- PUP 3D-GS: Principled Uncertainty Pruning for 3D Gaussian Splatting (CVPR 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.
