Unsupervised Cross-Domain Image Generation

Abstract

We study the problem of transferring a sample in one domain to an analog sample in another domain. Given two related domains, S and T, we would like to learn a generative function G that maps an input sample from S to the domain T, such that the output of a given function f, which accepts inputs in either domains, would remain unchanged. Other than the function f, the training data is unsupervised and consist of a set of samples from each domain. The Domain Transfer Network (DTN) we present employs a compound loss function that includes a multiclass GAN loss, an f-constancy component, and a regularizing component that encourages G to map samples from T to themselves. We apply our method to visual domains including digits and face images and demonstrate its ability to generate convincing novel images of previously unseen entities, while preserving their identity.

References in corpus (3)

• Conditional Generative Adversarial Nets
• Neural Photo Editing with Introspective Adversarial Networks
• Inverting face embeddings with convolutional neural networks

Cited by in corpus (27)

• CyCADA: Cycle-Consistent Adversarial Domain Adaptation
• Fader Networks: Manipulating Images by Sliding Attributes
• Domain Randomization for Transferring Deep Neural Networks from Simulation to the Real World
• Parallel-Data-Free Voice Conversion Using Cycle-Consistent Adversarial Networks
• Generating Diverse High-Fidelity Images with VQ-VAE-2
• Triangle Generative Adversarial Networks
• One-Sided Unsupervised Domain Mapping
• Label Efficient Learning of Transferable Representations across Domains and Tasks
• Minimal-Entropy Correlation Alignment for Unsupervised Deep Domain Adaptation
• 3D-A-Nets: 3D Deep Dense Descriptor for Volumetric Shapes with Adversarial Networks
• WaveCycleGAN2: Time-domain Neural Post-filter for Speech Waveform Generation
• VIGAN: Missing View Imputation with Generative Adversarial Networks
• GlyphGAN: Style-Consistent Font Generation Based on Generative Adversarial Networks
• Geometry-Aware Symmetric Domain Adaptation for Monocular Depth Estimation
• Learning Spatial Pyramid Attentive Pooling in Image Synthesis and Image-to-Image Translation
• Known-class Aware Self-ensemble for Open Set Domain Adaptation
• Sim-Real Joint Reinforcement Transfer for 3D Indoor Navigation
• Unsupervised Multi-Domain Image Translation with Domain-Specific Encoders/Decoders
• A Theory of Output-Side Unsupervised Domain Adaptation
• Generative Adversarial Network with Multi-Branch Discriminator for Cross-Species Image-to-Image Translation
• r-BTN: Cross-domain Face Composite and Synthesis from Limited Facial Patches
• Cross Domain Knowledge Learning with Dual-branch Adversarial Network for Vehicle Re-identification
• Image-to-Image Translation with Multi-Path Consistency Regularization
• Cross Domain Knowledge Transfer for Unsupervised Vehicle Re-identification
• Video-to-Video Translation for Visual Speech Synthesis
• Generative Adversarial Networks with Inverse Transformation Unit
• Learn to synthesize and synthesize to learn