Prune Wisely, Reconstruct Sharply: Compact 3D Gaussian Splatting via Adaptive Pruning and Difference-of-Gaussian Primitives

2026-02-27

AI Summary

This work addresses the challenge of excessive model size and limited scalability in 3D Gaussian Splatting (3DGS), which stems from its reliance on a large number of redundant Gaussians. To overcome this, the authors propose a reconstruction-aware adaptive pruning strategy coupled with a novel Difference-of-Gaussians primitive that models both positive and negative densities to enhance representational capacity. By dynamically optimizing the timing of pruning and the interval of refinement, the method achieves substantial model compression—reducing the number of Gaussians by up to 90%—while preserving or even improving rendering quality. This approach significantly advances the efficiency and compactness of 3DGS representations without compromising visual fidelity.

Technology Category

• Computer Vision: 3D Computer Vision
• Machine Learning: Deep Generative Models & Autoencoders
• Search and Optimization: Sampling/Simulation-based Search

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
• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines

Abstract

Recent significant advances in 3D scene representation have been driven by 3D Gaussian Splatting (3DGS), which has enabled real-time rendering with photorealistic quality. 3DGS often requires a large number of primitives to achieve high fidelity, leading to redundant representations and high resource consumption, thereby limiting its scalability for complex or large-scale scenes. Consequently, effective pruning strategies and more expressive primitives that can reduce redundancy while preserving visual quality are crucial for practical deployment. We propose an efficient, integrated reconstruction-aware pruning strategy that adaptively determines pruning timing and refining intervals based on reconstruction quality, thus reducing model size while enhancing rendering quality. Moreover, we introduce a 3D Difference-of-Gaussians primitive that jointly models both positive and negative densities in a single primitive, improving the expressiveness of Gaussians under compact configurations. Our method significantly improves model compactness, achieving up to 90\% reduction in Gaussian-count while delivering visual quality that is similar to, or in some cases better than, that produced by state-of-the-art methods. Code will be made publicly available.

Problem

• 3D Gaussian Splatting
• model compactness
• redundancy reduction
• scalability
• visual quality preservation

Innovation

• 3D Gaussian Splatting
• adaptive pruning
• Difference-of-Gaussians
• compact representation
• reconstruction-aware optimization