A Novel Benchmark and Dataset for Efficient 3D Gaussian Splatting with Gaussian Point Cloud Compression

TL;DR

This work tackles the storage burden of 3D Gaussian Splatting (3DGS) by introducing Gaussian point cloud geometry compression (GausPcgc) and a specialized training dataset (GausPcc-1K). It reframes Gaussian primitives as Gaussian point clouds and integrates AI-based point cloud compression into Gaussian pipelines, achieving reduced bitrates without sacrificing rendering quality. The authors present a Four-stage Occupancy Predictor and a dedicated multi-scale, occupancy-code–based compression scheme, trained on GausPcc-1K, and demonstrate clear bitrate improvements over traditional G-PCC baselines on Gaussian data (e.g., ~8.2% CR gain) while maintaining practical encoding/decoding speeds. This work establishes a first-of-its-kind Gaussian PCC benchmark and framework, enabling more efficient storage and faster rendering for applications in immersive media and autonomous systems, with future work on Gaussian attribute compression and duplicate-point handling.

Abstract

Recently, immersive media and autonomous driving applications have significantly advanced through 3D Gaussian Splatting (3DGS), which offers high-fidelity rendering and computational efficiency. Despite these advantages, 3DGS as a display-oriented representation requires substantial storage due to its numerous Gaussian attributes. Current compression methods have shown promising results but typically neglect the compression of Gaussian spatial positions, creating unnecessary bitstream overhead. We conceptualize Gaussian primitives as point clouds and propose leveraging point cloud compression techniques for more effective storage. AI-based point cloud compression demonstrates superior performance and faster inference compared to MPEG Geometry-based Point Cloud Compression (G-PCC). However, direct application of existing models to Gaussian compression may yield suboptimal results, as Gaussian point clouds tend to exhibit globally sparse yet locally dense geometric distributions that differ from conventional point cloud characteristics. To address these challenges, we introduce GausPcgc for Gaussian point cloud geometry compression along with a specialized training dataset GausPcc-1K. Our work pioneers the integration of AI-based point cloud compression into Gaussian compression pipelines, achieving superior compression ratios. The framework complements existing Gaussian compression methods while delivering significant performance improvements. All code, data, and pre-trained models will be publicly released to facilitate further research advances in this field.

Figures (13)

• Figure 1: Overview of our GausPcc-1K dataset and GausPcgc framework. Existing methods neglect Gaussian positions or use suboptimal G-PCC resulting in bitstream inefficiency. GausPcgc trained on our specialized dataset achieves superior inference speed and compression rates.
• Figure 2: Visualization of local density. We count neighbors within a 5×5×5 vicinity and render the results using a color gradient.
• Figure 3: Comparative analysis of local density and fractal dimension across different datasets.
• Figure 4: Introduction of the proposed GausPcc-1K Dataset.
• Figure 5: llustration of the proposed GausPcgc framework.