Reproduce “Look Both Ways Before You Cross: Lifting Cross Fields From 2D Visual Priors” 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
Dale Decatur, Jacob Serfaty, Oded Stein, Amir Vaxman, Rana Hanocka
CrossLift converts a triangle mesh into a quad mesh guided by visual features and surface geometry. It generates quad-grid images from six views, transfers their directions onto the surface, and combines per-view fields into a smooth cross field. On QuadWild300, it reports mean scaled Jacobian 0.9797 versus NeurCross’s 0.9737 and irregular-vertex percentage 1.4133% versus 1.6470%; semantic alignment is assessed qualitatively.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.