Exploring Explicit Domain Supervision for Latent Space Disentanglement in Unpaired Image-to-Image Translation

TL;DR

This paper introduces DosGAN, a novel unpaired image-to-image translation framework that explicitly uses domain supervision via a pre-trained classifier to improve translation quality and generalize to unseen domains.

Contribution

The paper proposes explicitly incorporating domain information through a pre-trained classifier, enhancing unpaired image translation and enabling transfer to unseen domains.

Findings

Effective in facial attribute, identity, and season translation

Able to transfer domain-specific features to unseen datasets

Achieves conditional translation between arbitrary image pairs

Abstract

Image-to-image translation tasks have been widely investigated with Generative Adversarial Networks (GANs). However, existing approaches are mostly designed in an unsupervised manner while little attention has been paid to domain information within unpaired data. In this paper, we treat domain information as explicit supervision and design an unpaired image-to-image translation framework, Domain-supervised GAN (DosGAN), which takes the first step towards the exploration of explicit domain supervision. In contrast to representing domain characteristics using different generators or domain codes, we pre-train a classification network to explicitly classify the domain of an image. After pre-training, this network is used to extract the domain-specific features of each image. Such features, together with the domain-independent features extracted by another encoder (shared across different domains), are used to generate image in target domain. Extensive experiments on multiple facial attribute translation, multiple identity translation, multiple season translation and conditional edges-to-shoes/handbags demonstrate the effectiveness of our method. In addition, we can transfer the domain-specific feature extractor obtained on the Facescrub dataset with domain supervision information to unseen domains, such as faces in the CelebA dataset. We also succeed in achieving conditional translation with any two images in CelebA, while previous models like StarGAN cannot handle this task.