Reproduce “RDT-3D: Resolution-Decoupled Training for High-Fidelity 3D Geometry 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

Jiale Zhang, Yang Liu, Chengyang Li, Jin Yuan, Qianxi Jia, Taicheng Huang

The authors introduce RDT-3D, training an image-to-3D refiner to use coarse positional guidance while learning fine geometry. RDT uses 512-resolution RoPE guidance and 1536-resolution supervision. On held-out Toys4K, it reports CD of 0.0164 and F-Score of 0.6886 versus UltraShape’s 0.0185 and 0.6676 after ICP alignment, within this benchmark evaluation.

Summary by alphaXiv.

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