Reversible GANs for Memory-Efficient Image-To-Image Translation

Tycho F.A. van der Ouderaa, Daniel E. Worrall; The IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2019, pp. 4720-4728

Abstract


The pix2pix and CycleGAN losses have vastly improved the qualitative and quantitative visual quality of results in image-to-image translation tasks. We extend this framework by exploring approximately invertible architectures which are well suited to these losses. These architectures are approximately invertible by design and thus partially satisfy cycle-consistency before training even begins. Furthermore, since invertible architectures have constant memory complexity in depth, these models can be built arbitrarily deep. We are able to demonstrate superior quantitative output on the Cityscapes and Maps datasets at near constant memory budget.

Related Material


[pdf] [supp]
[bibtex]
@InProceedings{Ouderaa_2019_CVPR,
author = {Ouderaa, Tycho F.A. van der and Worrall, Daniel E.},
title = {Reversible GANs for Memory-Efficient Image-To-Image Translation},
booktitle = {The IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
month = {June},
year = {2019}
}