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 (72)

• 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
• Unsupervised Domain Adaptation through Self-Supervision
• Generating Diverse High-Fidelity Images with VQ-VAE-2
• InstaGAN: Instance-aware Image-to-Image Translation
• Triangle Generative Adversarial Networks
• Label Efficient Learning of Transferable Representations across Domains and Tasks
• One-Sided Unsupervised Domain Mapping
• CalliGAN: Style and Structure-aware Chinese Calligraphy Character Generator
• Minimal-Entropy Correlation Alignment for Unsupervised Deep Domain Adaptation
• FaceLift: A transparent deep learning framework to beautify urban scenes
• BEV-Seg: Bird's Eye View Semantic Segmentation Using Geometry and Semantic Point Cloud
• 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
• Developing Personalized Models of Blood Pressure Estimation from Wearable Sensors Data Using Minimally-trained Domain Adversarial Neural Networks
• RL-CycleGAN: Reinforcement Learning Aware Simulation-To-Real
• GlyphGAN: Style-Consistent Font Generation Based on Generative Adversarial Networks
• SharinGAN: Combining Synthetic and Real Data for Unsupervised Geometry Estimation
• Geometry-Aware Symmetric Domain Adaptation for Monocular Depth Estimation
• Detection, Attribution and Localization of GAN Generated Images
• DG-Font: Deformable Generative Networks for Unsupervised Font Generation
• Image-to-image Translation via Hierarchical Style Disentanglement
• DART: Domain-Adversarial Residual-Transfer Networks for Unsupervised Cross-Domain Image Classification
• Learning Spatial Pyramid Attentive Pooling in Image Synthesis and Image-to-Image Translation
• Known-class Aware Self-ensemble for Open Set Domain Adaptation
• Few-shot Video-to-Video Synthesis
• Unsupervised Multi-Domain Image Translation with Domain-Specific Encoders/Decoders
• Sim-Real Joint Reinforcement Transfer for 3D Indoor Navigation
• Unsupervised Domain Adaptation for Visual Navigation
• CDCGen: Cross-Domain Conditional Generation via Normalizing Flows and Adversarial Training
• TSIT: A Simple and Versatile Framework for Image-to-Image Translation
• Multi-Target Domain Adaptation via Unsupervised Domain Classification for Weather Invariant Object Detection
• A Theory of Output-Side Unsupervised Domain Adaptation
• r-BTN: Cross-domain Face Composite and Synthesis from Limited Facial Patches
• Disentangled Cycle Consistency for Highly-realistic Virtual Try-On
• Learning Continuous Face Age Progression: A Pyramid of GANs
• Generative Adversarial Network with Multi-Branch Discriminator for Cross-Species Image-to-Image Translation
• Cross Domain Knowledge Learning with Dual-branch Adversarial Network for Vehicle Re-identification
• Unsupervised Vehicle Re-identification with Progressive Adaptation
• From Shadow Generation to Shadow Removal
• Learning Domain Adaptive Features with Unlabeled Domain Bridges
• Image-to-Image Translation with Multi-Path Consistency Regularization
• CADA: Multi-scale Collaborative Adversarial Domain Adaptation for Unsupervised Optic Disc and Cup Segmentation
• Learning a Domain-Agnostic Visual Representation for Autonomous Driving via Contrastive Loss
• Air-Writing Translater: A Novel Unsupervised Domain Adaptation Method for Inertia-Trajectory Translation of In-air Handwriting
• Cross Domain Knowledge Transfer for Unsupervised Vehicle Re-identification
• Sequential Gating Ensemble Network for Noise Robust Multi-Scale Face Restoration
• GANILLA: Generative Adversarial Networks for Image to Illustration Translation
• Auto-Encoding for Shared Cross Domain Feature Representation and Image-to-Image Translation
• Memory-guided Unsupervised Image-to-image Translation
• PSGAN++: Robust Detail-Preserving Makeup Transfer and Removal
• Generative Adversarial Networks with Inverse Transformation Unit
• Video-to-Video Translation for Visual Speech Synthesis
• Learn to synthesize and synthesize to learn
• Towards Visible and Thermal Drone Monitoring with Convolutional Neural Networks
• DGST : Discriminator Guided Scene Text detector
• Spectra2pix: Generating Nanostructure Images from Spectra
• Applications of Generative Adversarial Models in Visual Search Reformulation
• Towards Purely Unsupervised Disentanglement of Appearance and Shape for Person Images Generation
• MCMI: Multi-Cycle Image Translation with Mutual Information Constraints
• Head2HeadFS: Video-based Head Reenactment with Few-shot Learning
• Learning to Transfer Visual Effects from Videos to Images
• MakeupBag: Disentangling Makeup Extraction and Application
• Online Domain Adaptation for Continuous Cross-Subject Liver Viability Evaluation Based on Irregular Thermal Data
• Localising In Complex Scenes Using Balanced Adversarial Adaptation
• Enlisting 3D Crop Models and GANs for More Data Efficient and Generalizable Fruit Detection
• Profile to Frontal Face Recognition in the Wild Using Coupled Conditional GAN
• A Survey on Adversarial Image Synthesis
• Multi-Channel Auto-Encoders and a Novel Dataset for Learning Domain Invariant Representations of Histopathology Images