[论文] Towards Practical Compression of 3D Gaussian Splatting

论文概要

研究领域: CV 作者: Pengpeng Yu, Yueru Chen, Fei Song, Tai Qin, Qi Zhang, Jing Wang, Yulan Guo 发布时间: 2026-09-24 arXiv: 2609.30245

中文摘要

3D 高斯泼溅（3DGS）可实现高质量的新视角合成，但需要大量存储空间。现有的压缩方法通常依赖对不规则 3D 表示的空间上下文建模，增加了训练和编码的复杂度。同时，浮点上下文推理可能引入跨平台的数值不一致性，导致熵解码失败。为解决这些实际挑战，我们提出 COSA-GS，通过锚点级因果分解构建上下文，避免空间聚合。具体而言，我们利用每个锚点坐标派生的几何上下文来建模一个紧凑的可学习锚点潜变量，然后将该锚点潜变量与几何上下文融合形成用于属性编码的锚点上下文。所得上下文模型仅由线性变换和激活函数组成，架构简洁。我们使用率-失真优化结合自适应高斯剪枝来训练 COSA-GS。此外，我们还为上下文模型开发了量化感知训练和整数推理，以实现跨平台熵解码符号的比特级一致。实验表明，COSA-GS 达到了最先进的压缩性能，同时保持了快速且一致的跨平台解码，为实用的 3DGS 压缩提供了一个简洁而有效的框架。代码见 https://github.com/pengpeng-yu/COSA-GS。

原文摘要

3D Gaussian Splatting (3DGS) enables high-quality novel-view synthesis but requires substantial storage. Existing compression methods often rely on spatial context modeling over irregular 3D representations, increasing the complexity of training and coding. Meanwhile, floating-point context inference can introduce numerical inconsistencies across platforms, causing entropy-decoding failures. To address these practical challenges, we propose COSA-GS, which constructs context without spatial aggregation through anchor-wise causal factorization. Specifically, we use geometry context derived from each anchor's coordinates to model a compact learnable anchor latent. The anchor latent is then fused with the geometry context to form an anchor context for attribute coding. The resulting context mo...

#论文 #arXiv #CV #小凯