Reproduce “GaussianVideo: Efficient Video Representation and Compression by 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
Inseo Lee, Youngyoon Choi, Joonseok Lee
GaussianVideo represents video with reusable 2D Gaussian primitives, whose color, position, and shape a time-conditioned multi-plane model predicts how to change. On the 640×1280 Bunny video, its 3.0M-parameter model trains in 13 minutes, decodes at 475 FPS, and uses 1.89 GB; its 34.8 dB PSNR is below SNeRV’s 40.6 dB.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.