Compression method cuts 3D gaussian splatting data size by 30 percent
Compressing 3D Gaussian Splatting via Cross-Representation Priors
Computer Vision and Pattern Recognition
Summary
3D Gaussian Splatting creates detailed 3D scenes but uses a lot of storage space because it stores many small pieces called Gaussian primitives. The authors propose a new way to shrink this data by better organizing these pieces based on their features, not just their positions, allowing some pieces to help describe others. They also gather and share common scene details globally, so each piece doesn’t have to store the same information repeatedly. Their approach reduces file size by about 30% while keeping the quality of the 3D images similar.
What this means in practice
- •For 3d graphics engineers: Reduce data size of 3D Gaussian Splatting scenes for efficient storage and transmission without losing visual quality.
- •For game developers: Enable faster loading and lower bandwidth use for detailed 3D environments using compressed Gaussian Splatting representations.
Authors
Yezheng Zhang, Huanxiong Liang, Chuqin Zhou, Guo Lu, Wenjun Zhang
Abstract
3D Gaussian Splatting (3DGS) enables high-quality novel view synthesis but incurs high storage and transmission costs due to dense Gaussian primitives. Recent anchor-based compression reduces per-primitive redundancy, yet redundancy across anchors remains largely unexploited. We propose CRP-GS (Cross-Representation Priors for Gaussian Splatting), a rate-distortion optimized compression framework that leverages cross-representation priors to improve anchor-level entropy modeling. First, a Correspondence-Oriented Hierarchical Structure (COHS) organizes anchors by feature correspondence rather than spatial proximity, constructing root-leaf dependencies so that selected anchors can act as informative priors to conditionally encode others, yielding more accurate likelihood prediction and lower conditional entropy. Second, Shared Feature Aggregation (SFA) extracts globally shared features from a contextual hash grid and injects them into anchor representations, factoring out scene-consistent low-frequency information that would otherwise be redundantly embedded in individual anchors. Both modules are trained under a unified rate-distortion objective to balance bitrate reduction and rendering fidelity. Experiments across multiple benchmarks show that CRP-GS achieves a favorable overall rate-distortion trade-off, yielding around 30% average bitrate reduction compared to anchor-based baselines while maintaining comparable rendering quality.