Abstract
We propose a new approach to human clothing modeling based on point clouds.Within this approach, we learn a deep model that can predict point clouds ofvarious outfits, for various human poses and for various human body shapes.Notably, outfits of various types and topologies can be handled by the samemodel. Using the learned model, we can infer geometry of new outfits from aslittle as a singe image, and perform outfit retargeting to new bodies in newposes. We complement our geometric model with appearance modeling that uses thepoint cloud geometry as a geometric scaffolding, and employs neural point-basedgraphics to capture outfit appearance from videos and to re-render the capturedoutfits. We validate both geometric modeling and appearance modeling aspects ofthe proposed approach against recently proposed methods, and establish theviability of point-based clothing modeling.