TryOnGAN: Body-aware Try-on via Layered Interpolation

Katie Lewis
Srivatsan Varadharajan
SIGGRAPH (2021)

Abstract

Given a pair of images—target person and garment on another person—we automatically generate the target person in the given garment. Previous methods mostly focused on texture transfer via paired data training, while overlooking body shape deformations, skin color, and seamless blending of garment with the person. This work focuses on those three components, while also not requiring paired data training. We designed a pose conditioned StyleGAN2 architecture with a clothing segmentation branch that is trained on images of people wearing garments. Once trained, we propose a new layered latent space interpolation method that allows us to preserve and synthesize skin color and target body shape while transferring the garment from a different person. We demonstrate results on high resolution 512x512 images, and extensively compare to state of the art in try-on on both latent space generated and real images.

Research Areas