Research map / 3D vision / Gaussian splatting
RobustSplat: Decoupling Densification and Dynamics for Transient-Free 3DGS
Chuanyu Fu, Yuqi Zhang, Kunbin Yao, Guanying Chen, Yuan Xiong, Chuan Huang, Shuguang Cui, Xiaochun Cao
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 transient mask MLP to 3DGS (ablation) | 3DGS · NeRF On-the-go Spot | PSNR: 18.54 -> 24.75 | Improved |
| add delayed Gaussian growth to 3DGS+Mask+MB (ablation) | RobustSplat · NeRF On-the-go Spot | PSNR: 25.25 -> 26.21 | Improved |
| add scale-cascaded mask bootstrapping to 3DGS+Mask+DG (ablation) | RobustSplat · NeRF On-the-go Spot | PSNR: 25.61 -> 26.21 | Improved |
| add delayed Gaussian growth to plain 3DGS (no mask) (ablation) | 3DGS · NeRF On-the-go Spot | PSNR: 18.54 -> 23.42 | Improved |
| add mask regularization (ablation) | RobustSplat · NeRF On-the-go Spot | PSNR: 25.63 -> 26.21 | Improved |
| add image residual loss to mask supervision (ablation) | RobustSplat · NeRF On-the-go mean | PSNR: 23.77 -> 24.54 | Improved |
| add feature cosine similarity loss to mask supervision (ablation) | RobustSplat · NeRF On-the-go mean | PSNR: 23.46 -> 24.54 | Improved |
| use DINOv2 features instead of SAM2 for mask MLP input (ablation) | RobustSplat · NeRF On-the-go mean | PSNR: 24.01 -> 24.54 | Improved |
| use DINOv2 features instead of Stable Diffusion for mask MLP input (ablation) | RobustSplat · NeRF On-the-go mean | PSNR: 24.04 -> 24.54 | Improved |
| RobustSplat versus T-3DGS on NeRF On-the-go | T-3DGS · NeRF On-the-go mean | PSNR: 22.87 -> 23.22 | Improved |
| RobustSplat versus T-3DGS on RobustNeRF | T-3DGS · RobustNeRF mean | PSNR: 28.25 -> 29.36 | Improved |
| RobustSplat versus T-3DGS on On-the-go II | T-3DGS · NeRF On-the-go II mean | PSNR: 22.63 -> 23.02 | 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.
- WildGaussians — WildGaussians: 3D Gaussian Splatting In the Wild (NeurIPS 2024) · better on 4 of 4 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
- MaskGaussian: Adaptive 3D Gaussian Representation from Probabilistic Masks (CVPR 2025 · similarity 0.92)
- Speedy-Splat: Fast 3D Gaussian Splatting with Sparse Pixels and Sparse Primitives (CVPR 2025 · similarity 0.92)
- Decoupling Appearance Variations with 3D Consistent Features in Gaussian Splatting (AAAI 2025 · similarity 0.91)
- Robust Neural Rendering in the Wild with Asymmetric Dual 3D Gaussian Splatting (NeurIPS 2025 · similarity 0.91)
- LP-3DGS: Learning to Prune 3D Gaussian Splatting (NeurIPS 2024 · similarity 0.91)
- PUP 3D-GS: Principled Uncertainty Pruning for 3D Gaussian Splatting (CVPR 2025 · 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.
