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

TL;DR

This paper introduces an adaptive pruning strategy and a novel Difference-of-Gaussians primitive for 3D Gaussian Splatting, significantly reducing model size while maintaining or improving rendering quality.

Contribution

It presents a reconstruction-aware pruning method and a new primitive type that enhance compactness and expressiveness of 3D Gaussian Splatting models.

Findings

Achieves up to 90% reduction in Gaussian primitives

Maintains or improves visual quality compared to state-of-the-art

Enhances model efficiency and scalability for large scenes

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.