Research map / 3D vision / Gaussian splatting
BA-GS: Bayesian Adaptive Gaussian Splatting for SFM-Free 3D Reconstruction
Zhongjie Ma, Di Lin, Xin Wang, Haotian Dong, Chong Wang, Dongdong Wu, Changqing Zhang
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 VB-GMM initialization clustering (ablation) | BA-GS · Tanks and Temples 12-view | PSNR: 31.0429 -> 31.6129 | Improved |
| add prior-guided adaptive density control (ablation) | BA-GS · Tanks and Temples 12-view | PSNR: 31.0182 -> 31.6129 | Improved |
| add Adaptive Kalman position filtering (ablation) | BA-GS · Tanks and Temples 12-view | PSNR: 31.1920 -> 31.6129 | Improved |
| add all BA-GS components to InstantSplat baseline (ablation) | InstantSplat (MASt3R) · Tanks and Temples 12-view | PSNR: 30.0851 -> 31.6129 | Improved |
| add VB-GMM pruning of redundant primitives (ablation) | BA-GS · Tanks and Temples 12-view | rendering time (s): 190.25 -> 153.88 | Improved |
| use BA-GS with VGGT initialization instead of InstantSplat | VGGT initialization · Tanks and Temples 12-view | PSNR: 27.7603 -> 29.9707 | Improved |
| use BA-GS with VGGT initialization instead of InstantSplat | VGGT initialization · LLFF 18-view | PSNR: 18.9155 -> 24.7906 | Improved |
| add BA-GS to MASt3R InstantSplat at 3 views | MASt3R InstantSplat · Tanks and Temples 3-view | PSNR: 23.5750 -> 23.5667 | No clear effect |
| add BA-GS to MASt3R InstantSplat with more views | MASt3R InstantSplat · MVImgNet 18-view | PSNR: 23.4952 -> 25.0474 | Improved |
| replace MASt3R with VGGT initialization in InstantSplat | InstantSplat · Tanks and Temples 12-view | PSNR: 30.0851 -> 27.7603 | Hurt |
Closest papers by meaning
- Metropolis-Hastings Sampling for 3D Gaussian Reconstruction (NeurIPS 2025 · similarity 0.92)
- Variational Multi-scale Representation for Estimating Uncertainty in 3D Gaussian Splatting (NeurIPS 2024 · similarity 0.91)
- 4D Gaussian Splatting in the Wild with Uncertainty-Aware Regularization (NeurIPS 2024 · similarity 0.90)
- PUP 3D-GS: Principled Uncertainty Pruning for 3D Gaussian Splatting (CVPR 2025 · similarity 0.90)
- SfM-Free 3D Gaussian Splatting via Hierarchical Training (CVPR 2025 · similarity 0.90)
- Structure Consistent Gaussian Splatting with Matching Prior for Few-shot Novel View Synthesis (NeurIPS 2024 · 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.
