Reproduce “SplatFlux: Feed-Forward 4D Gaussian Reconstruction and Generation from Video Latents” 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

Runmin Zhang, Binglin Ren, Zhu Yu, Fudong Wang, Jiangke Lin, Yichun Shentu, Si-Yuan Cao, Yi Yuan and 1 more

SplatFlux maps captured video or generated video latents into renderable 4D Gaussian scenes. It predicts which primitives are dynamic, aggregates static primitives across frames, and moves dynamic ones using local motion. This supports rendering from different viewpoints and times. On NVIDIA Dynamic Scene, it reports 21.56 PSNR versus 18.83 for NeoVerse; motion between timestamps uses linear interpolation.

Summary by alphaXiv.

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