Reproduce “HNeRV: A Hybrid Neural Representation for Videos” 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
Hao Chen, Matt Gwilliam, Ser-Nam Lim, Abhinav Shrivastava
The authors introduce HNeRV, which encodes each frame into a compact, content-adaptive embedding for a video-specific decoder that reconstructs frames independently. Its decoder shifts capacity toward later, higher-resolution stages. On Bunny after 300 epochs, mean PSNR across four matched model sizes is 33.90 dB for HNeRV, 31.90 for E-NeRV, and 29.88 for NeRV; this is a specific regression comparison, and fitting a new video still requires training.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.