Reproduce “Deep Lambertian Networks” 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
Yichuan Tang, Ruslan Salakhutdinov, Geoffrey Hinton
Deep Lambertian Networks (DLNs) estimate illumination-stable face appearance and geometry from images to support recognition under changing light. They combine learned priors over albedo and surface normals with a Lambertian image-formation model and an explicit light variable. In the one-training-image Yale B comparison, plotted test error is below four baselines; this benchmark does not establish superiority or identify the cause.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.