Research map / 3D vision / Gaussian splatting
Improving Explicit Dynamic Gaussian Splatting Optimization via Update Mixture
Renjie Ding, Yaonan Wang, Min Liu, Jialin Zhu, Jiazheng Wang, Jiahao Zhao, Xiao Tan, Feixiang He, Xiang Chen
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 Strictly Sparse Update (ablation) | 4DGS · N3DV Cut Roasted Beef | PSNR: 32.239 -> 33.141 | Improved |
| add regularization (RSR) to SSU (ablation) | 4DGS · N3DV Cut Roasted Beef | PSNR: 33.141 -> 33.820 | Improved |
| add regularization (RSR) to SSU (ablation) | 4DGS · N3DV Flame Salmon | PSNR: 29.786 -> 29.715 | No clear effect |
| add estimated epsilon (ablation) | 4DGS · N3DV Cut Roasted Beef | PSNR: 33.820 -> 34.202 | Improved |
| add SAA (ablation) | 4DGS · N3DV Cut Roasted Beef | PSNR: 34.202 -> 34.321 | Improved |
| use batch size 1 instead of 4 with full method (ablation) | 4DGS · N3DV Cut Roasted Beef | PSNR: 34.321 -> 33.617 | Hurt |
| apply SSU and regularization with the 4DGS epsilon 1e-15 (ablation) | Ex4DGS · N3DV | PSNR: 31.315 -> 31.267 | No clear effect |
| raise epsilon from 1e-8 to 1e-6 with SSU (ablation) | Ex4DGS · N3DV | PSNR: 31.315 -> 31.618 | Improved |
| use estimated epsilon instead of fixed 1e-6 or 1e-7 (ablation) | Ex4DGS · N3DV | PSNR: 31.618 -> 31.729 | Improved |
| add RSR with estimated epsilon (ablation) | Ex4DGS · N3DV | PSNR: 31.729 -> 32.112 | Improved |
| add SAA (ablation) | Ex4DGS · N3DV | PSNR: 32.159 -> 32.204 | Improved |
| add full update mixture | Ex4DGS · Technicolor | PSNR: 31.809 -> 32.281 | Improved |
Closest papers by meaning
- Efficient Gaussian Splatting for Monocular Dynamic Scene Rendering via Sparse Time-Variant Attribute Modeling (AAAI 2025 · similarity 0.92)
- Speedy-Splat: Fast 3D Gaussian Splatting with Sparse Pixels and Sparse Primitives (CVPR 2025 · similarity 0.91)
- Fully Explicit Dynamic Gaussian Splatting (NeurIPS 2024 · similarity 0.91)
- GaussianSpa: An "Optimizing-Sparsifying" Simplification Framework for Compact and High-Quality 3D Gaussian Splatting (CVPR 2025 · similarity 0.91)
- SC-GS: Sparse-Controlled Gaussian Splatting for Editable Dynamic Scenes (CVPR 2024 · similarity 0.91)
- Swift4D: Adaptive divide-and-conquer Gaussian Splatting for compact and efficient reconstruction of dynamic scene (ICLR 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.
