ML Anthology

SharinGAN: Combining Synthetic and Real Data for Unsupervised Geometry Estimation

Koutilya Pnvr, Hao Zhou, David Jacobs
CVPR 2020
doi:10.1109/CVPR42600.2020.01399 /cvpr/2020/pnvr2020cvpr-sharingan/

Abstract

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.

PDF CVPR Semantic Scholar

Cite

Text

Pnvr et al. "SharinGAN: Combining Synthetic and Real Data for Unsupervised Geometry Estimation." Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2020. doi:10.1109/CVPR42600.2020.01399

Markdown

[Pnvr et al. "SharinGAN: Combining Synthetic and Real Data for Unsupervised Geometry Estimation." Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2020.](https://mlanthology.org/cvpr/2020/pnvr2020cvpr-sharingan/) doi:10.1109/CVPR42600.2020.01399

BibTeX

@inproceedings{pnvr2020cvpr-sharingan,
  title     = {{SharinGAN: Combining Synthetic and Real Data for Unsupervised Geometry Estimation}},
  author    = {Pnvr, Koutilya and Zhou, Hao and Jacobs, David},
  booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
  year      = {2020},
  doi       = {10.1109/CVPR42600.2020.01399},
  url       = {https://mlanthology.org/cvpr/2020/pnvr2020cvpr-sharingan/}
}