Reproduce “3DRAG: 3D-Native Object Drag Editing with Generative Models” 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
Kuixiang Shao, Yuyang Jiao, Chenghao Zhang, Wenbo Zhao, Yinuo Bai, Jingyi Yu
3DRAG performs point-and-drag editing of 3D objects by moving content within a generative model. It maps user-specified displacements through an occupancy-aware voxel grid and transfers cached source features to destination locations while preserving regions outside the edit. On its 3D benchmark, it reports a Dragging Accuracy Index of 0.187 versus 0.368 for APAP, with lower scores better. Its sparse 64³ grid limits fine detail.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.