Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in Coakley et alK. L. Coakley, T. Snelleman, H. Hoos, and O. E. Gundersen, "The embrace of open science: An analysis of a decade of AI research and 56 800 conference papers," Under Review, 2026..
Reframing Gaussian Splatting Densification with Complexity-Density Consistency of Primitives
Authors: Zhemeng Dong, Junjun Jiang, Youyu Chen, Jiaxin Zhang, Kui Jiang, Xianming Liu
NeurIPS 2025 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Extensive experiments demonstrate that our CDC-GS surpasses the baseline methods in rendering quality by a large margin using the same amount of Gaussians. And we provide insightful analysis to reveal that our method serves perpendicularly to rendering loss in guiding Gaussian primitive allocation. |
| Researcher Affiliation | Academia | Zhemeng Dong, Junjun Jiang , Youyu Chen, Jiaxin Zhang, Kui Jiang, Xianming Liu Faculty of Computing, Harbin Institute of Technology |
| Pseudocode | Yes | Algorithm 1 provides a detailed pseudocode of the training pipeline in CDC-GS. |
| Open Source Code | No | We use publicly available datasets described in Section4.1, and the code will be released in the future. |
| Open Datasets | Yes | Dataset. We evaluate our method on the real-world datasets widely adopted in novel view synthesis. Specifically, we follow the standard protocol in vanilla 3DGS [1], using 9 scenes from the Mip Ne RF360 [30] dataset, the playroom and drjohnson scenes from Deep Blending [52] dataset, and the train and truck scenes from Tanks & Temples [51] dataset. |
| Dataset Splits | No | All metrics are computed on test views using the original evaluation protocol. |
| Hardware Specification | Yes | All experiments are performed on a single NVIDIA RTX 3090 GPU. |
| Software Dependencies | No | Our method is implemented on top of the official 3DGS codebase [1]. |
| Experiment Setup | Yes | Our complexity-aware densification threshold is configured with τlow = 1.5 10 4, τupp = 2 10 4 and λ = 6. Other hyper-parameters are consistent with the vanilla 3DGS [1]. |