ZstGAN: An Adversarial Approach for Unsupervised Zero-Shot Image-to-Image Translation
This addresses a limitation in image-to-image translation for scenarios with unseen domains, though it is incremental as it builds on existing adversarial and zero-shot learning frameworks.
The paper tackles the problem of unsupervised zero-shot image-to-image translation, where models must translate images between domains not seen during training, and proposes ZstGAN, which achieves significant accuracy improvements over state-of-the-art methods on CUB and FLO datasets.
Image-to-image translation models have shown remarkable ability on transferring images among different domains. Most of existing work follows the setting that the source domain and target domain keep the same at training and inference phases, which cannot be generalized to the scenarios for translating an image from an unseen domain to another unseen domain. In this work, we propose the Unsupervised Zero-Shot Image-to-image Translation (UZSIT) problem, which aims to learn a model that can translate samples from image domains that are not observed during training. Accordingly, we propose a framework called ZstGAN: By introducing an adversarial training scheme, ZstGAN learns to model each domain with domain-specific feature distribution that is semantically consistent on vision and attribute modalities. Then the domain-invariant features are disentangled with an shared encoder for image generation. We carry out extensive experiments on CUB and FLO datasets, and the results demonstrate the effectiveness of proposed method on UZSIT task. Moreover, ZstGAN shows significant accuracy improvements over state-of-the-art zero-shot learning methods on CUB and FLO.