Reproduce “SAFE-Hair: Scalp-Anchored Fields for Exportable Single-View Hair Reconstruction” 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

Yucheng Wang, Zedong Wang, Yuetong Wu, Yue Ma, Dan Xu

SAFE-Hair reconstructs 3D hair from a single portrait by conditioning a scalp-based generative model on image features and integrating strand curves from fixed follicles with persistent identities. On HairBench-3D, it reports 3.28 mm Chamfer distance, 81.7 F-score, 69.8 invisible-region F-score, 3.9% penetration, and 99.8% valid roots in the benchmark, though single-view ambiguity and downstream cleanup remain.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.