用扩散模型实现3D高斯泼溅的千倍压缩，保持高质量渲染。
Nix and Fix: Targeting 1000x Compression of 3D Gaussian Splatting with Diffusion Models
- 通过感知噪声引导的单步蒸馏，从压缩数据中恢复细节。
- 在0.1MB极低码率下仍保持最优视觉质量，压缩比达1000倍。
- 适合需要超低存储/传输成本的3D内容部署场景。
3D高斯泼溅（3DGS）革新了新视角渲染技术，不再依赖密集空间点进行隐式推断，而是采用稀疏高斯分布，实现实时性能，但显著增加存储需求，限制了速率受限应用。3DGS压缩成为研究热点以缓解此问题。尽管进展显著，但在低码率下仍会引入明显伪影，严重影响视觉质量。本文提出NiFi方法，通过基于扩散模型的一步式蒸馏实现极端压缩，具备伪影感知能力。实验表明，该方法在极低码率（低至0.1MB）下仍达到当前最优感知质量，相较原始3DGS实现接近1000倍的压缩率提升，且在相同感知表现下大幅降低数据量。代码已开源。
原文摘要 · Abstract (English)
3D Gaussian Splatting (3DGS) revolutionized novel view rendering. Instead of inferring from dense spatial points, as implicit representations do, 3DGS uses sparse Gaussians. This enables real-time performance but increases space requirements, hindering rate-constrained applications. 3DGS compression emerged as a field aimed at alleviating this issue. While impressive progress has been made, at low rates, compression introduces artifacts that degrade visual quality significantly. We introduce NiFi, a method for extreme 3DGS compression through restoration via artifact-aware, diffusion-based one-step distillation. We show that our method achieves state-of-the-art perceptual quality at extremely low rates, down to 0.1 MB, and towards 1000x rate improvement over 3DGS at comparable perceptual performance. Code is available at: https://github.com/ceteke/nifi
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。