Big GANs Are Watching You: Towards Unsupervised Object Segmentation with Off-the-Shelf Generative Models
Andrey Sergeevich Voynov, Stanislav Morozov, Artem Babenko · arXiv (Cornell University) · 2021
Since collecting pixel-level groundtruth data is expensive, unsupervised visual understanding problems are currently an active research topic. In particular, several recent methods based on generative models have achieved promising results for object segmentation and saliency detection. However, since generative models are known to be unstable and sensitive to hyperparameters, the training of these methods can be challenging and time-consuming. In this work, we introduce an alternative, much simpler way to exploit generative models for unsupervised object segmentation. First, we explore the latent spaces of the publicly available unsupervised models, such as BigBiGAN and StyleGAN2, and reveal the ``segmenting'' latent directions that can be used to obtain saliency masks for GAN-produced images. These masks then are used to train a discriminative segmentation model. Being very simple and easy-to-reproduce, our approach outperforms the state-of-the-art on common benchmarks in the unsupervised scenario. All the code and the pretrained models are available online.