3D inpainting of Garment models by cGAN

Post date: 02-Feb-2018 00:35:42

Description

In this research, we will investigate a conditional GAN (cGAN) to tackle the problem of 3D model synthesis. This is a generalised version of image inpainting, where the key problem is completing missing (or corrupted) 3D geometry as well as its texture information. A potential application of this research is producing a more realistic 3D garment model, which could improve the UX of the garment interactions in future VA/AR.

Project period

This project will be funded by University of Portsmouth Global PhD Scholarship and commence in Oct/2018 (see PhD advertisement).