Reproduce “NiLBS: Neural Inverse Linear Blend Skinning” 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
Timothy Jeruzalski, David I.W. Levin, Alec Jacobson, Paul Lalonde, Mohammad Norouzi, Andrea Tagliasacchi
NiLBS proposes reusing a rest-pose occupancy field across character poses. It learns to map queries in posed space back to rest coordinates, where the cached field can be queried. A multilayer perceptron predicts skinning weights, while a ghost-bone channel is intended to handle background queries. A 2D example illustrates an occupancy artifact addressed by the correction; empirical evaluation remains future work.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.