Improving Densification in 3D Gaussian Splatting for High-Fidelity Rendering

TL;DR

This paper improves the densification process in 3D Gaussian Splatting, enhancing rendering quality and fidelity by introducing novel strategies for Gaussian splitting, overfitting mitigation, and geometric distortion reduction.

Contribution

It presents a comprehensive densification pipeline with new techniques like Edge-Aware Score and Long-Axis Split, significantly improving 3DGS performance without extra overhead.

Findings

Achieves state-of-the-art rendering quality with fewer Gaussians.

Reduces geometric distortions in Gaussian splitting.

Enhances densification effectiveness without additional training or inference costs.

Abstract

Although 3D Gaussian Splatting (3DGS) has achieved impressive performance in real-time rendering, its densification strategy often results in suboptimal reconstruction quality. In this work, we present a comprehensive improvement to the densification pipeline of 3DGS from three perspectives: when to densify, how to densify, and how to mitigate overfitting. Specifically, we propose an Edge-Aware Score to effectively select candidate Gaussians for splitting. We further introduce a Long-Axis Split strategy that reduces geometric distortions introduced by clone and split operations. To address overfitting, we design a set of techniques, including Recovery-Aware Pruning, Multi-step Update, and Growth Control. Our method enhances rendering fidelity without introducing additional training or inference overhead, achieving state-of-the-art performance with fewer Gaussians.