Exploring Unlabeled Faces For Novel Attribute Discovery

Despite remarkable success in unpaired image-to-image translation, existing approaches still require a large amount of labeled images. This is a bottleneck against their realworld applications; in practice, a model trained on a labeled dataset, such as CelebA dataset, does not work well for test images from a different distribution - limiting their applications to unlabeled images of a much larger quantity. In this paper; we attempt to alleviate this necessity for labeled data in the facial image translation domain. We aim to explore the degree to which you can discover novel attributes from unlabeled faces and perform high-quality translation. Ib this end, we use prior knowledge about the visual world as guidance to discover novel attributes and transfer them via a novel normalization method. Experiments show that our method trained on unlabeled data produces high-quality translations, preserves identity, and is perceptually realistic, as good as, or better than, state-of-the-art methods trained on labeled data.