Reproduce “Visual-to-Executable Procedural Reconstruction of Buildings” 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
Zichong Zhu, Zhihao Liu, Andrei Sharf, Zhanglin Cheng
The method converts a building-facade photo into an editable procedural asset by predicting facade structure and component attributes, then compiling them into CGA rules. By iteration 2, window, floor, and door count accuracy reaches 0.80, 0.86, and 0.81, while TRC falls from 4.32 to 1.47. Unseen sides are completed for instantiation, not inferred from the photo.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.