Reproduce “Canonical Factors for Hybrid Neural Fields” 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
Brent Yi, Weijia Zeng, Sam Buchanan, Yi Ma
TILTED learns rotations that map scene coordinates into frames where factored neural fields can represent scenes efficiently, while retaining existing factors and decoder. On Lego, axis-aligned K-Planes PSNR falls from 35.31 ± 0.02 to 33.29 ± 0.11 under random 3D rotation. On 18 real-world scenes, it improves training PSNR over same-size axis-aligned models, but evidence is limited to tested settings and training metrics.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.