EqGAN - Improving GAN Equilibrium by Raising Spatial Awareness
Improving GAN Equilibrium by Raising Spatial Awareness
Jianyuan Wang, Ceyuan Yang, Yinghao Xu, Yujun Shen, Hongdong Li, Bolei Zhou
arXiv preprint
[Paper] [Project Page] [Demo]
In Generative Adversarial Networks (GANs), a generator (G) and a discriminator (D) are expected to reach a certain equilibrium where D cannot distinguish the generated images from the real ones. However, in practice it is difficult to achieve such an equilibrium in GAN training, instead, D almost always surpasses G. We attribute this phenomenon to the information asymmetry that D learns its own visual attention when determining whether an image is real or fake, but G has no explicit clue on which regions to focus on.
To alleviate the issue of D dominating the competition in GANs, we aim to raise the spatial awareness of G. We encode randomly sampled multi-level heatmaps into the intermediate layers of G as an inductive bias. We further propose to align the spatial awareness of G with the attention map induced from D. Through this way we effectively lessen the information gap between D and G. Extensive results show that our method pushes the two-player game in GANs closer to the equilibrium, leading to a better synthesis performance. As a byproduct, the introduced spatial awareness facilitates interactive editing over the output synthesis.
BibTeX
@article{wang2021eqgan,
title = {Improving GAN Equilibrium by Raising Spatial Awareness},
author = {Wang, Jianyuan and Yang, Ceyuan and Xu, Yinghao and Shen, Yujun and Li, Hongdong and Zhou, Bolei},
article = {arXiv preprint},
year = {2021}
}
66 Nov 28, 2022
438 Dec 29, 2022
34 Jan 19, 2022
70 Jan 2, 2023
151 Dec 28, 2022
![[PyTorch] Official implementation of CVPR2021 paper](https://github.com/XuyangBai/PointDSC/raw/master/figures/pipeline.png)
153 Dec 14, 2022
97 Dec 20, 2022
482 Dec 18, 2022
4 Apr 28, 2022
136 Dec 18, 2022
124 Dec 30, 2022
101 Jan 1, 2023
82 Sep 15, 2022
4 Jul 27, 2022
191 Dec 31, 2022
18 Dec 23, 2022
28 Nov 25, 2022
46 Dec 14, 2022