Class-Splitting Generative Adversarial Networks
This work addresses the challenge of enhancing GAN performance for image generation tasks, offering a method that can be applied even without labeled data, though it is incremental as it builds on existing conditional GAN frameworks.
The paper tackles the problem of improving sample quality in conditional Generative Adversarial Networks (GANs) by augmenting class labels through clustering in the learned representation space, achieving state-of-the-art Inception scores on CIFAR-10 and STL-10 datasets in both supervised and unsupervised setups.
Generative Adversarial Networks (GANs) produce systematically better quality samples when class label information is provided., i.e. in the conditional GAN setup. This is still observed for the recently proposed Wasserstein GAN formulation which stabilized adversarial training and allows considering high capacity network architectures such as ResNet. In this work we show how to boost conditional GAN by augmenting available class labels. The new classes come from clustering in the representation space learned by the same GAN model. The proposed strategy is also feasible when no class information is available, i.e. in the unsupervised setup. Our generated samples reach state-of-the-art Inception scores for CIFAR-10 and STL-10 datasets in both supervised and unsupervised setup.