Reproduce “GETA-3DGS: Automatic Joint Structured Pruning and Quantization for 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

Baobing Zhang, Wanxin Sui

GETA-3DGS compresses 3DGS scenes by jointly pruning whole Gaussian primitives and assigning attribute-specific bit widths, guided by rendered effects. On the nine-scene Mip-NeRF 360 test average, it reduces storage from 20.83 MB to 4.86 MB, while PSNR falls from 25.15 dB to 23.36 dB. Uniform six-bit quantization loses 2.18 dB on average, and nominal storage budgets do not reliably control final size.

Summary by alphaXiv.

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