NanoGS: Training-Free Gaussian Splat Simplification

2026-03-16

AI Summary

This work addresses the high storage and transmission costs of 3D Gaussian Splatting (3DGS) models, which typically contain millions of primitives, and the impracticality of existing compression methods that rely on time-consuming post-training optimization and image-based supervision. We propose the first lightweight simplification framework that operates without training or rendering supervision, leveraging only geometric and statistical properties to pairwise merge Gaussians within local neighborhoods via a sparse spatial graph. Our approach employs moment-matching approximation under mass conservation, a mixture-distribution discrepancy–driven merging cost, and efficient candidate selection, enabling fast CPU execution. It substantially reduces primitive count while preserving high-quality novel view synthesis and remains fully compatible with the standard 3DGS parameterization, thereby enhancing model compactness and deployment practicality.

Technology Category

• Computer Vision: 3D Computer Vision
• Machine Learning: Learning on the Edge & Model Compression
• Data Mining & Knowledge Management: Data Compression

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

3D Gaussian Splat (3DGS) enables high-fidelity, real-time novel view synthesis by representing scenes with large sets of anisotropic primitives, but often requires millions of Splats, incurring significant storage and transmission costs. Most existing compression methods rely on GPU-intensive post-training optimization with calibrated images, limiting practical deployment. We introduce NanoGS, a training-free and lightweight framework for Gaussian Splat simplification. Instead of relying on image-based rendering supervision, NanoGS formulates simplification as local pairwise merging over a sparse spatial graph. The method approximates a pair of Gaussians with a single primitive using mass preserved moment matching and evaluates merge quality through a principled merge cost between the original mixture and its approximation. By restricting merge candidates to local neighborhoods and selecting compatible pairs efficiently, NanoGS produces compact Gaussian representations while preserving scene structure and appearance. NanoGS operates directly on existing Gaussian Splat models, runs efficiently on CPU, and preserves the standard 3DGS parameterization, enabling seamless integration with existing rendering pipelines. Experiments demonstrate that NanoGS substantially reduces primitive count while maintaining high rendering fidelity, providing an efficient and practical solution for Gaussian Splat simplification. Our project website is available at https://saliteta.github.io/NanoGS/.

Problem

• 3D Gaussian Splat
• compression
• storage cost
• transmission cost
• real-time rendering

Innovation

• training-free
• Gaussian Splat simplification
• local pairwise merging
• moment matching
• lightweight compression