Self-labeled Conditional GANs
This work addresses the problem of training conditional GANs without human-annotated labels, which is significant for researchers and practitioners working with large-scale datasets where fine-grained annotations are scarce.
This paper proposes a self-labeled conditional GAN framework that automatically obtains labels from data. The generator outperforms unconditional GANs in FID on ImageNet and LSUN, and also surpasses class-conditional GANs trained with human labels on CIFAR10 and CIFAR100.
This paper introduces a novel and fully unsupervised framework for conditional GAN training in which labels are automatically obtained from data. We incorporate a clustering network into the standard conditional GAN framework that plays against the discriminator. With the generator, it aims to find a shared structured mapping for associating pseudo-labels with the real and fake images. Our generator outperforms unconditional GANs in terms of FID with significant margins on large scale datasets like ImageNet and LSUN. It also outperforms class conditional GANs trained on human labels on CIFAR10 and CIFAR100 where fine-grained annotations or a large number of samples per class are not available. Additionally, our clustering network exceeds the state-of-the-art on CIFAR100 clustering.