Reproduce “UniCross3D: Unified Cross-View and Cross-Domain Diffusion for Consistent Single-Image 3D 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
U-Chae Jun, Jaeeun Ko, Jiwoo Kang
UniCross3D generates paired color and surface-normal images from one input, then uses them to guide NeuS reconstruction. It trains the diffusion model with a per-view penalty on bottleneck-feature variation across noise draws and a contrastive loss aligning matched color–normal embeddings. On 500 Google Scanned Objects inputs, its Chamfer Distance is 0.0125 versus Era3D’s 0.0153, supporting better results in this evaluation, not uniquely correct recovery of unseen shapes.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.