Reproduce “MIKU: Material Invariance and Knowledge Unification for PBR Generation” 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
Yajie Bao, Jiaye Fu, Youtian Lin, Shuang Wu, Jiachen Qian, Feihu Zhang, Xun Cao, Yao Yao
MIKU generates geometry-aware multi-view albedo and metallic–roughness maps after standardizing artist-authored material parameters using category-level ranges. The authors test whether this reduces prediction variation, reporting a consistency mean squared error of 0.0258 versus 0.0513 with original materials under semantic-preserving image edits. This supports greater stability in that test, not physically measured material accuracy.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.