We propose a novel method for combining synthetic and real images when training networks to determine geometric information from a single image. We suggest a method for mapping both image types into a single, shared domain. This is connected to a primary network for end-to-end training. Ideally, this results in images from two domains that present shared information to the primary network. Our experiments demonstrate significant improvements over the state-of-the-art in two important domains, surface normal estimation of human faces and monocular depth estimation for outdoor scenes, both in an unsupervised setting.
@article{arxiv.2006.04026,
title = {SharinGAN: Combining Synthetic and Real Data for Unsupervised Geometry Estimation},
author = {Koutilya PNVR and Hao Zhou and David Jacobs},
journal= {arXiv preprint arXiv:2006.04026},
year = {2020}
}
Comments
Accepted to CVPR 2020. Supplementary material added towards the end instead of a separate file. A Github link to the code is also provided in this submission