Reproduce “EntropyGS: An Efficient Entropy Coding on 3D Gaussian Splatting” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Yuning Huang, Jiahao Pang, Fengqing Zhu, Dong Tian

EntropyGS compresses pre-generated 3D Gaussian Splatting models using attribute-specific distributions and entropy coding, rather than a learned probability model. It quantizes attributes according to their sensitivity, then codes them with arithmetic coding or G-PCC. On Mip-NeRF 360, its medium setting reaches a 28.25× compression ratio with a −0.04 dB change in PSNR; comparisons cover compression-specific methods, not joint-optimization methods.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.