Diverse Image Generation via Self-Conditioned GANs
This addresses mode collapse for GAN users, though it appears incremental as it builds on existing class-conditional GAN frameworks with a novel clustering approach.
The paper tackles the problem of mode collapse in GANs by introducing an unsupervised method that conditions generation on automatically discovered cluster labels from discriminator features, resulting in improved image diversity and quality metrics on standard benchmarks and large-scale datasets.
We introduce a simple but effective unsupervised method for generating realistic and diverse images. We train a class-conditional GAN model without using manually annotated class labels. Instead, our model is conditional on labels automatically derived from clustering in the discriminator's feature space. Our clustering step automatically discovers diverse modes, and explicitly requires the generator to cover them. Experiments on standard mode collapse benchmarks show that our method outperforms several competing methods when addressing mode collapse. Our method also performs well on large-scale datasets such as ImageNet and Places365, improving both image diversity and standard quality metrics, compared to previous methods.