FairStyle: Debiasing StyleGAN2 with Style Channel Manipulations
This addresses fairness issues in generative models for applications requiring diverse and unbiased image generation, representing an incremental improvement over existing debiasing techniques.
The paper tackles the problem of bias in StyleGAN2-generated images by proposing a method to directly modify a pre-trained model to produce balanced sets of images with respect to attributes like gender and eyeglasses, achieving debiasing within minutes without quality loss.
Recent advances in generative adversarial networks have shown that it is possible to generate high-resolution and hyperrealistic images. However, the images produced by GANs are only as fair and representative as the datasets on which they are trained. In this paper, we propose a method for directly modifying a pre-trained StyleGAN2 model that can be used to generate a balanced set of images with respect to one (e.g., eyeglasses) or more attributes (e.g., gender and eyeglasses). Our method takes advantage of the style space of the StyleGAN2 model to perform disentangled control of the target attributes to be debiased. Our method does not require training additional models and directly debiases the GAN model, paving the way for its use in various downstream applications. Our experiments show that our method successfully debiases the GAN model within a few minutes without compromising the quality of the generated images. To promote fair generative models, we share the code and debiased models at http://catlab-team.github.io/fairstyle.