CityGaussianV2: Efficient and Geometrically Accurate Reconstruction for Large-Scale Scenes

TL;DR

CityGaussianV2 introduces a scalable, efficient method for large-scale scene reconstruction that improves geometric accuracy and reduces computational costs by innovative densification, depth regression, and parallel training techniques.

Contribution

It presents novel techniques for large-scale scene reconstruction, enhancing geometric accuracy and efficiency over previous Gaussian Splatting methods.

Findings

Achieves up to 10× compression in training data

Reduces training time by at least 25%

Decreases memory usage by 50%

Abstract

Recently, 3D Gaussian Splatting (3DGS) has revolutionized radiance field reconstruction, manifesting efficient and high-fidelity novel view synthesis. However, accurately representing surfaces, especially in large and complex scenarios, remains a significant challenge due to the unstructured nature of 3DGS. In this paper, we present CityGaussianV2, a novel approach for large-scale scene reconstruction that addresses critical challenges related to geometric accuracy and efficiency. Building on the favorable generalization capabilities of 2D Gaussian Splatting (2DGS), we address its convergence and scalability issues. Specifically, we implement a decomposed-gradient-based densification and depth regression technique to eliminate blurry artifacts and accelerate convergence. To scale up, we introduce an elongation filter that mitigates Gaussian count explosion caused by 2DGS degeneration. Furthermore, we optimize the CityGaussian pipeline for parallel training, achieving up to 10$\times$ compression, at least 25% savings in training time, and a 50% decrease in memory usage. We also established standard geometry benchmarks under large-scale scenes. Experimental results demonstrate that our method strikes a promising balance between visual quality, geometric accuracy, as well as storage and training costs. The project page is available at https://dekuliutesla.github.io/CityGaussianV2/.