Label-Removed Generative Adversarial Networks Incorporating with K-Means
This work addresses the challenge of expensive label acquisition in GANs for image generation, offering an incremental improvement by adapting existing methods to reduce labeling needs.
The authors tackled the problem of reducing dependence on labeled data in generative adversarial networks (GANs) by proposing an unconditional model, KM-GAN, which incorporates K-Means clustering into the discriminator's features, resulting in sample quality comparable to some conditional models on datasets like MNIST and CIFAR-10.
Generative Adversarial Networks (GANs) have achieved great success in generating realistic images. Most of these are conditional models, although acquisition of class labels is expensive and time-consuming in practice. To reduce the dependence on labeled data, we propose an un-conditional generative adversarial model, called K-Means-GAN (KM-GAN), which incorporates the idea of updating centers in K-Means into GANs. Specifically, we redesign the framework of GANs by applying K-Means on the features extracted from the discriminator. With obtained labels from K-Means, we propose new objective functions from the perspective of deep metric learning (DML). Distinct from previous works, the discriminator is treated as a feature extractor rather than a classifier in KM-GAN, meanwhile utilization of K-Means makes features of the discriminator more representative. Experiments are conducted on various datasets, such as MNIST, Fashion-10, CIFAR-10 and CelebA, and show that the quality of samples generated by KM-GAN is comparable to some conditional generative adversarial models.