Reproduce “GeoGSplat: Geometry-Grounded Non-Pixel-Aligned 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

Jiale Wang, Chen Zhao, Haoyang Zhang, Yanhao Wu, Fei He, Yanhu Shan, Rui Wu, Wei Ke and 1 more

GeoGSplat predicts renderable 3D scenes from posed images using 65,536 Gaussians, independent of input-view count and image resolution. It uses spatially structured Anchors and Triplanes with geometric supervision to guide point placement. On NRGBD with four input views, it reports 24.78 dB PSNR and 0.017 CD, versus TokenGS’s 23.82 dB and 0.277; these are benchmark results, not speed measurements.

Summary by alphaXiv.

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