Robust Conditional GAN from Uncertainty-Aware Pairwise Comparisons
This addresses the annotation bottleneck for conditional GANs in image editing, though it is an incremental improvement over existing methods.
The paper tackles the problem of conditional GANs requiring large annotations by proposing PC-GAN, which uses weak supervision from pairwise comparisons for image attribute editing, achieving performance comparable to fully-supervised methods and outperforming unsupervised baselines.
Conditional generative adversarial networks have shown exceptional generation performance over the past few years. However, they require large numbers of annotations. To address this problem, we propose a novel generative adversarial network utilizing weak supervision in the form of pairwise comparisons (PC-GAN) for image attribute editing. In the light of Bayesian uncertainty estimation and noise-tolerant adversarial training, PC-GAN can estimate attribute rating efficiently and demonstrate robust performance in noise resistance. Through extensive experiments, we show both qualitatively and quantitatively that PC-GAN performs comparably with fully-supervised methods and outperforms unsupervised baselines.