Stabilizing GANs with Soft Octave Convolutions

Ricard Durall, Ricard Durall, Ricard Durall, Franz-Josef Pfreundt, Janis Keuper, Janis Keuper

2021

Abstract

Motivated by recently published methods using frequency decompositions of convolutions (e.g. Octave Convolutions), we propose a novel convolution scheme to stabilize the training and reduce the likelihood of a mode collapse. The basic idea of our approach is to split convolutional filters into additive high and low frequency parts, while shifting weight updates from low to high during the training. Intuitively, this method forces GANs to learn low frequency coarse image structures before descending into fine (high frequency) details. We also show, that the use of the proposed soft octave convolutions reduces common artifacts in the frequency domain of generated images. Our approach is orthogonal and complementary to existing stabilization methods and can simply be plugged into any CNN based GAN architecture. Experiments on the CelebA dataset show the effectiveness of the proposed method.

Download


Paper Citation


in Harvard Style

Durall R., Pfreundt F. and Keuper J. (2021). Stabilizing GANs with Soft Octave Convolutions. In Proceedings of the 16th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2021) - Volume 4: VISAPP; ISBN 978-989-758-488-6, SciTePress, pages 15-23. DOI: 10.5220/0010178700150023


in Bibtex Style

@conference{visapp21,
author={Ricard Durall and Franz-Josef Pfreundt and Janis Keuper},
title={Stabilizing GANs with Soft Octave Convolutions},
booktitle={Proceedings of the 16th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2021) - Volume 4: VISAPP},
year={2021},
pages={15-23},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0010178700150023},
isbn={978-989-758-488-6},
}


in EndNote Style

TY - CONF

JO - Proceedings of the 16th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2021) - Volume 4: VISAPP
TI - Stabilizing GANs with Soft Octave Convolutions
SN - 978-989-758-488-6
AU - Durall R.
AU - Pfreundt F.
AU - Keuper J.
PY - 2021
SP - 15
EP - 23
DO - 10.5220/0010178700150023
PB - SciTePress