Research map / 3D vision / Gaussian splatting
Reframing Gaussian Splatting Densification with Complexity-Density Consistency of Primitives
Zhemeng Dong, Junjun Jiang, Youyu Chen, Jiaxin Zhang, Kui Jiang, Xianming Liu
Read the paperCheck your idea against 137,217 papers
AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Proposes a densification strategy for 3D Gaussian Splatting that allocates primitives by enforcing consistency between scene complexity and primitive density, instead of relying only on rendering-loss-driven under-reconstruction signals.
Problem
Standard 3DGS densification is driven by rendering loss, which is dominated by low-frequency regions and under-models high-frequency texture-rich details.
Approach
Reframes densification around complexity-density consistency: primitive allocation is guided so that local scene complexity (e.g., high-frequency content) matches the density of Gaussian primitives, rather than marking under-reconstructed regions by loss magnitude.
Experimental setup
Builds on: 3D Gaussian Splatting
Closest papers by meaning
- F⁴Splat: Feed-Forward Predictive Densification for Feed-Forward 3D Gaussian Splatting (ECCV 2026 · similarity 0.93)
- Optimized Minimal 3D Gaussian Splatting (NeurIPS 2025 · similarity 0.93)
- EDGS: Eliminating Densification for Efficient Convergence of 3DGS (CVPR 2026 · similarity 0.93)
- Gaussian Herding across Pens: An Optimal Transport Perspective on Global Gaussian Reduction for 3DGS (NeurIPS 2025 · similarity 0.93)
- DashGaussian: Optimizing 3D Gaussian Splatting in 200 Seconds (CVPR 2025 · similarity 0.92)
- Speedy-Splat: Fast 3D Gaussian Splatting with Sparse Pixels and Sparse Primitives (CVPR 2025 · similarity 0.92)
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.
