🚀 NeuralGS: Bridging Neural Fields and 3D Gaussian Splatting for Compact 3D Representation

arXiv Code(coming soon)

Zhenyu Tang1*, Chaoran Feng1*, Xinhua Cheng1, Wangbo Yu1, Junwu Zhang1,
Yuan Liu2†, Xiaoxiao Long2, Wenping Wang3, Li Yuan 1†,
1Peking University
2Hong Kong University of Science and Technology
3Texas A&M University
*These authors contributed equally to this work.
†Corresponding author.

We propose NeuralGS, a novel framework that effectively adopts the neural field representation to encode the attributes of 3D Gaussians with multiple tiny MLPs, only requiring a small model size even for a large-scale scene.

đź“Ś Abstract


3D Gaussian Splatting (3DGS) achieves impressive quality and rendering speed, but with millions of 3D Gaussians and significant storage and transmission costs. In this paper, we aim to develop a simple yet effective method called NeuralGS that compresses the original 3DGS into a compact representation. Our observation is that neural fields like NeRF can represent complex 3D scenes with Multi-Layer Perceptron (MLP) neural networks using only a few megabytes. Thus, NeuralGS effectively adopts the neural field representation to encode the attributes of 3D Gaussians with MLPs, only requiring a small storage size even for a large-scale scene. To achieve this, we adopt a clustering strategy and fit the Gaussians within each cluster using different tiny MLPs, based on importance scores of Gaussians as fitting weights. We experiment on multiple datasets, achieving a 91× average model size reduction without harming the visual quality.

đź’ˇ Compression Method

We propose a simple yet effective framework NeuralGS, which adopts the neural field representation to encode the attributes of 3D Gaussians with MLPs to enable a compact 3D representation. Specifically, as shown in Figure 1, we use a designed criterion to assess the importance of each Gaussian, allowing us to prune Gaussians that have minimal impact on renderings. To reduce variations among Gaussians, we cluster these 3D Gaussians based on their attributes, ensuring similarity within each cluster. Then, each cluster is then assigned a tiny MLP that fits the attributes of its 3D Gaussians. Given the varying contributions of each Gaussian to the renderings, we apply Gaussian’s importances as the fitting weights in the MLP fitting. We also incorporate a fine-tuning stage along with frequency loss to restore quality and preserve high-frequency details. We address the storage and rendering issue of 3D Gaussian Splatting by compressing the reconstructed scene parameters and rendering the compressed representation via GPU rasterization. To compress the scenes, we first analyze its components and observe that the Spherical Harmonics (SH) coefficients and the multivariate Gaussian parameters take up the majority of storage space and are highly redundant. Our compression pipeline consists of three steps:


Compression Pipeline

Figure 2: Overview of NeuralGS. (A) In Sec. 3.2, for each Gaussian GSj in the scene, we first calculate its global importance score Sj (Eq. 1) and prune unimportant Gaussians. (B) In Sec. 3.3, we cluster the pruned Gaussians and use different tiny MLPs to map the positions to Gaussian attributes of different clusters with the loss (Eq. 3) using the importance score as weights. (C) In Sec. 3.4, we fine-tune the tiny MLPs of all clusters with photorealistic loss (Eq. 4) and frequency loss (Eq. 5) to restore quality.

đź”— Results


Qualitative Comparisons

Compression Pipeline
Figure 2: Qualitative comprison on the mentioned four datasets.

Quantitative Comparisons

Table 1. Quantitative results evaluated on Mip-NeRF 360, Tanks&Temples, and Deep Blending datasets. We highlight the best-performing results in bold and the second-best results in underline for all compression methods.
Compression Pipeline

Table 2. Quantitative results of the proposed method evaluated on the NeRF-Synthetic dataset. We highlight the best-performing results best-performing results in bold and the second-best results in underline for all compression methods.
Table 2

🙋‍♂️ Conclusion

In this paper, we introduce NeuralGS, a novel and effective post-compression approach for 3D Gaussian splatting. The core of our approach lies in leveraging compact neural fields to encode the attributes of 3D Gaussians with MLPs, significantly reducing the memory requirements of 3DGS. Thus, we design multiple neural fields based on clusters and incorporate importance scores as fitting weights to enhance the fittting quality of Gaussian attributes. Additionally, we introduce frequency loss during the fine-tuning stage to further preserve high-frequency details. Extensive experiments demonstrate that our method achieves comparable or even superior performance to existing compression methods while utilizing less model size. Overall, NeuralGS paves the way for directly compressing original 3DGS with neural fields.

BibTeX

@article{tang2025neuralgs,
            title={NeuralGS: Bridging Neural Fields and 3D Gaussian Splatting for Compact 3D Representations},
            author={Tang*, Zhenyu and Feng*, Chaoran and Cheng, Xinhua and Yu, Wangbo and Zhang, Junwu and Liu, Yuan and Long, Xiaoxiao and Wang, Wenping and Yuan, Li},
            journal={arXiv preprint arXiv:2503.23162},
            year={2025}
            }