Reproduce “SoftPAT3D: Physics-Augmented Text-to-3D Scene Generation with Deformables” 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
Lei Shu, Alberta Longhini, Leonidas Guibas, Gordon Wetzstein, Jeannette Bohg
SoftPAT3D prepares generated cloth for room simulation by constructing class-based rest shapes, repairing occluded supports, placing cloth above the prompt-named part, and settling deformable and rigid bodies with a shared contact model. On a matched 34-scene test, it completes 21 scenes, compared with none for the same-asset rigid pipeline. Templates replace generated size and folds with class priors.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.