GANerated Hands for Real-Time 3D Hand Tracking from Monocular RGB

Spotlight @ Computer Vision and Pattern Recognition (CVPR) 2018, Salt Lake City, Utah, USA

Download Video(MP4, 720p, 93.4 MB)

Abstract

We address the highly challenging problem of real-time 3D hand tracking based on a monocular RGB-only sequence. Our tracking method combines a convolutional neural network with a kinematic 3D hand model, such that it generalizes well to unseen data, is robust to occlusions and varying camera viewpoints, and leads to anatomically plausible as well as temporally smooth hand motions. For training our CNN we propose a novel approach for the synthetic generation of training data that is based on a geometrically consistent image-to-image translation network. To be more specific, we use a neural network that translates synthetic images to “real” images, such that the so-generated images follow the same statistical distribution as real-world hand images. For training this translation network we combine an adversarial loss and a cycle-consistency loss with a geometric consistency loss in order to preserve geometric properties (such as hand pose) during translation. We demonstrate that our hand tracking system outperforms the current state-of-the-art on challenging RGB-only footage.

Downloads

For acquiring access to the CNN model, please send an e-mail from your institutional mail address stating your full name and affiliation.

Citation

BibTeX, 1 KB

@inproceedings{GANeratedHands_CVPR2018,
 author = {Mueller, Franziska and Bernard, Florian and Sotnychenko, Oleksandr and Mehta, Dushyant and Sridhar, Srinath and Casas, Dan and Theobalt, Christian},
 title = {GANerated Hands for Real-Time 3D Hand Tracking from Monocular RGB},
 booktitle = {Proceedings of Computer Vision and Pattern Recognition ({CVPR})},
 url = {https://handtracker.mpi-inf.mpg.de/projects/GANeratedHands/},
 numpages = {11},
 month = {June},
 year = {2018}
} 
	

Acknowledgments

This research was funded by the ERC Starting Grant projects CapReal (335545).
Dan Casas was supported by a Marie Curie Individual Fellow, grant agreement 707326.

Contact

Franziska Mueller
frmueller@mpi-inf.mpg.de

This page is Zotero and Mendeley translator friendly. Page last updated 28-March-2018.

Imprint/Impressum | Data Protection/Datenschutzhinweis