Unsupervised Image-to-Image Translation Networks
It addresses the problem of learning joint distributions across image domains without paired data, which is useful for applications like style transfer and domain adaptation, but is incremental as it builds on existing GAN-based methods.
The paper tackles unsupervised image-to-image translation by proposing a framework based on Coupled GANs with a shared-latent space assumption, achieving high-quality results on tasks like street scene, animal, and face translation and state-of-the-art performance in domain adaptation.
Unsupervised image-to-image translation aims at learning a joint distribution of images in different domains by using images from the marginal distributions in individual domains. Since there exists an infinite set of joint distributions that can arrive the given marginal distributions, one could infer nothing about the joint distribution from the marginal distributions without additional assumptions. To address the problem, we make a shared-latent space assumption and propose an unsupervised image-to-image translation framework based on Coupled GANs. We compare the proposed framework with competing approaches and present high quality image translation results on various challenging unsupervised image translation tasks, including street scene image translation, animal image translation, and face image translation. We also apply the proposed framework to domain adaptation and achieve state-of-the-art performance on benchmark datasets. Code and additional results are available in https://github.com/mingyuliutw/unit .