EfficientGS: Streamlining Gaussian Splatting for Large-Scale High-Resolution Scene Representation

TL;DR

EfficientGS significantly improves large-scale, high-resolution 3D scene representation by optimizing Gaussian management, reducing computational costs, and maintaining high fidelity, enabling faster training and rendering with smaller models.

Contribution

We introduce EfficientGS, a novel method that optimizes Gaussian splatting for large-scale scenes through selective Gaussian management, pruning, and sparse SH integration.

Findings

Reduces training and rendering times substantially.

Achieves tenfold smaller model size compared to traditional 3DGS.

Maintains high rendering fidelity on 4K+ datasets.

Abstract

In the domain of 3D scene representation, 3D Gaussian Splatting (3DGS) has emerged as a pivotal technology. However, its application to large-scale, high-resolution scenes (exceeding 4k$\times$4k pixels) is hindered by the excessive computational requirements for managing a large number of Gaussians. Addressing this, we introduce 'EfficientGS', an advanced approach that optimizes 3DGS for high-resolution, large-scale scenes. We analyze the densification process in 3DGS and identify areas of Gaussian over-proliferation. We propose a selective strategy, limiting Gaussian increase to key primitives, thereby enhancing the representational efficiency. Additionally, we develop a pruning mechanism to remove redundant Gaussians, those that are merely auxiliary to adjacent ones. For further enhancement, we integrate a sparse order increment for Spherical Harmonics (SH), designed to alleviate storage constraints and reduce training overhead. Our empirical evaluations, conducted on a range of datasets including extensive 4K+ aerial images, demonstrate that 'EfficientGS' not only expedites training and rendering times but also achieves this with a model size approximately tenfold smaller than conventional 3DGS while maintaining high rendering fidelity.