# SafeguardGS: 3D Gaussian Primitive Pruning While Avoiding Catastrophic Scene Destruction arXiv: https://arxiv.org/abs/2405.17793 ยท HTML: https://mlsanity.com/paper/2405.17793/ ## Key facts - Cited by 7 papers in the map (data through 14 Oct 2026). - Builds on 41 of its 59 references that are in the map. - First posted on arXiv 28 May 2024. ## Builds on (most cited first, 30 of 41) - [3D Gaussian Splatting for Real-Time Radiance Field Rendering](https://mlsanity.com/paper/2308.04079/) (2023) - [Instant Neural Graphics Primitives with a Multiresolution Hash Encoding](https://mlsanity.com/paper/2201.05989/) (2022) - [Mip-NeRF 360: Unbounded Anti-Aliased Neural Radiance Fields](https://mlsanity.com/paper/2111.12077/) (2021) - [Mip-NeRF: A Multiscale Representation for Anti-Aliasing Neural Radiance Fields](https://mlsanity.com/paper/2103.13415/) (2021) - [Plenoxels: Radiance Fields without Neural Networks](https://mlsanity.com/paper/2112.05131/) (2021) - [pixelNeRF: Neural Radiance Fields from One or Few Images](https://mlsanity.com/paper/2012.02190/) (2020) - [TensoRF: Tensorial Radiance Fields](https://mlsanity.com/paper/2203.09517/) (2022) - [NeRF in the Wild: Neural Radiance Fields for Unconstrained Photo Collections](https://mlsanity.com/paper/2008.02268/) (2020) - [Direct Voxel Grid Optimization: Super-fast Convergence for Radiance Fields Reconstruction](https://mlsanity.com/paper/2111.11215/) (2021) - [PlenOctrees for Real-time Rendering of Neural Radiance Fields](https://mlsanity.com/paper/2103.14024/) (2021) - [NeRF++: Analyzing and Improving Neural Radiance Fields](https://mlsanity.com/paper/2010.07492/) (2020) - [GIRAFFE: Representing Scenes as Compositional Generative Neural Feature Fields](https://mlsanity.com/paper/2011.12100/) (2020) - [Neural Scene Flow Fields for Space-Time View Synthesis of Dynamic Scenes](https://mlsanity.com/paper/2011.13084/) (2020) - [IBRNet: Learning Multi-View Image-Based Rendering](https://mlsanity.com/paper/2102.13090/) (2021) - [GRAF: Generative Radiance Fields for 3D-Aware Image Synthesis](https://mlsanity.com/paper/2007.02442/) (2020) - [StyleNeRF: A Style-based 3D-Aware Generator for High-resolution Image Synthesis](https://mlsanity.com/paper/2110.08985/) (2021) - [Baking Neural Radiance Fields for Real-Time View Synthesis](https://mlsanity.com/paper/2103.14645/) (2021) - [Neuralangelo: High-Fidelity Neural Surface Reconstruction](https://mlsanity.com/paper/2306.03092/) (2023) - [LightGaussian: Unbounded 3D Gaussian Compression with 15x Reduction and 200+ FPS](https://mlsanity.com/paper/2311.17245/) (2023) - [Dynamic View Synthesis from Dynamic Monocular Video](https://mlsanity.com/paper/2105.06468/) (2021) - [INeRF: Inverting Neural Radiance Fields for Pose Estimation](https://mlsanity.com/paper/2012.05877/) (2020) - [Compact 3D Gaussian Representation for Radiance Field](https://mlsanity.com/paper/2311.13681/) (2023) - [NeuS2: Fast Learning of Neural Implicit Surfaces for Multi-view Reconstruction](https://mlsanity.com/paper/2212.05231/) (2022) - [MobileNeRF: Exploiting the Polygon Rasterization Pipeline for Efficient Neural Field Rendering on Mobile Architectures](https://mlsanity.com/paper/2208.00277/) (2022) - [DONeRF: Towards Real-Time Rendering of Compact Neural Radiance Fields using Depth Oracle Networks](https://mlsanity.com/paper/2103.03231/) (2021) - [Mini-Splatting: Representing Scenes with a Constrained Number of Gaussians](https://mlsanity.com/paper/2403.14166/) (2024) - [NeRF-SLAM: Real-Time Dense Monocular SLAM with Neural Radiance Fields](https://mlsanity.com/paper/2210.13641/) (2022) - [Deblur-NeRF: Neural Radiance Fields from Blurry Images](https://mlsanity.com/paper/2111.14292/) (2021) - [NICER-SLAM: Neural Implicit Scene Encoding for RGB SLAM](https://mlsanity.com/paper/2302.03594/) (2023) - [SPIn-NeRF: Multiview Segmentation and Perceptual Inpainting with Neural Radiance Fields](https://mlsanity.com/paper/2211.12254/) (2022) ## Built on by (most cited first, 7 of 7) - [Improving 3D Gaussian Splatting Compression by Scene-Adaptive Lattice Vector Quantization](https://mlsanity.com/paper/2509.13482/) (2025) - [POTR: Post-Training 3DGS Compression](https://mlsanity.com/paper/2601.14821/) (2026) - [ControlGS: Consistent Structural Compression Control for Deployment-Aware Gaussian Splatting](https://mlsanity.com/paper/2505.10473/) (2025) - [Matryoshka Gaussian Splatting](https://mlsanity.com/paper/2603.19234/) (2026) - [MesonGS++: Post-training Compression of 3D Gaussian Splatting with Hyperparameter Searching](https://mlsanity.com/paper/2604.26799/) (2026) - [CaT-GS: Efficient 3DGS Rendering for Large Scale Scenes via Inter-frame Caching and Tile Scheduling](https://mlsanity.com/paper/2607.17842/) (2026) - [LiteTex-GS: Fast and Lightweight Texturing for Gaussian Splatting](https://mlsanity.com/paper/2609.23380/) (2026) ## Abstract (arXiv) 3D Gaussian Splatting (3DGS) has made significant strides in novel view synthesis. However, its suboptimal densification process results in the excessively large number of Gaussian primitives, which impacts frame-per-second and increases memory usage, making it unsuitable for low-end devices. To address this issue, many follow-up studies have proposed various pruning techniques with score functions designed to identify and remove less important primitives. Nonetheless, a comprehensive discussion of their effectiveness and implications across all techniques is missing. In this paper, we are the first to categorize 3DGS pruning techniques into two types: Scene-level pruning and Pixel-level pruning, distinguished by their scope for ranking primitives. Our subsequent experiments reveal that, while scene-level pruning leads to disastrous quality drops under extreme decimation of Gaussian primitives, pixel-level pruning not only sustains relatively high rendering quality with minuscule performance degradation but also provides an inherent boundary of pruning, i.e., a safeguard of Gaussian pruning. Building on this observation, we further propose multiple variations of score functions based on the factors of rendering equations and discover that assessing based on color similarity with blending weight is the most effective method for discriminating insignificant primitives. In our experiments, our SafeguardGS with the optimal score function shows the highest PSNR-per-primitive performance under an extreme pruning setting, retaining only about 10% of the primitives from the original 3DGS scene (i.e., 10x compression ratio). We believe our research provides valuable insights for optimizing 3DGS for future works.