Learning to Discover Cross-Domain Relations with Generative Adversarial Networks

2017/03/15 by Taeksoo Kim, Taek‐Soo Kim, Moonsu Cha +9 · 2 voices · 732 citations
Computer Science · Mathematics · #Adversarial system #Artificial intelligence #Computer science #Computer security #Domain (mathematical analysis) #Face (sociological concept) #Generative Adversarial Networks and Image Synthesis #Generative grammar #Identity (music) #Key (lock) #Mathematics #Music and Audio Processing #Pairing #Task (project management) #Theoretical computer science #Video Analysis and Summarization #cs.CV

paper · pdf · doi:10.48550/arxiv.1703.05192

published in arXiv (Cornell University), 1857-1865 (Cornell University) · Accepted to International Conference on Machine Learning (ICML) 2017

openalex publication_date 2017/03/15 · arxiv published 2017/03/15 · arxiv created 2017/05/15 · arxiv updated 2017/05/16 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/05

Abstract

While humans easily recognize relations between data from different domains without any supervision, learning to automatically discover them is in general very challenging and needs many ground-truth pairs that illustrate the relations. To avoid costly pairing, we address the task of discovering cross-domain relations given unpaired data. We propose a method based on generative adversarial networks that learns to discover relations between different domains (DiscoGAN). Using the discovered relations, our proposed network successfully transfers style from one domain to another while preserving key attributes such as orientation and face identity. Source code for official implementation is publicly available https://github.com/SKTBrain/DiscoGAN

Citations

• Auto-Encoding Variational Bayes
• Improved Techniques for Training GANs
• Mode Regularized Generative Adversarial Networks
• Generative Visual Manipulation on the Natural Image Manifold
• Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks

Cited by

• Plasticine: A Traceable Diffusion Model for Medical Image Translation
• High-dimensional Mean-Field Games by Particle-based Flow Matching
• Structure-Preserving Unpaired Image Translation to Photometrically Calibrate JunoCam with Hubble Data
• TReFT: Taming Rectified Flow Models For One-Step Image Translation
• Recursive Chaining of Reversible Image-to-image Translators For Face\n Aging
• Detecting GAN generated Fake Images using Co-occurrence Matrices
• Adversarial Transfer Learning for Cross-domain Visual Recognition
• Diverse Semantic Image Synthesis via Probability Distribution Modeling
• Unsupervised Contrastive Photo-to-Caricature Translation based on Auto-distortion
• Generative Adversarial Networks for Image and Video Synthesis: Algorithms and Applications
• Open-Edit: Open-Domain Image Manipulation with Open-Vocabulary Instructions
• Smoothing the Disentangled Latent Style Space for Unsupervised Image-to-Image Translation
• Augmenting Images for ASR and TTS through Single-loop and Dual-loop Multimodal Chain Framework
• Learning Cross-domain Generalizable Features by Representation Disentanglement
• Training End-to-end Single Image Generators without GANs
• Dual Contrastive Learning for Unsupervised Image-to-Image Translation
• Improving Shape Deformation in Unsupervised Image-to-Image Translation
• Translating and Segmenting Multimodal Medical Volumes with Cycle- and Shape-Consistency Generative Adversarial Network
• StarGAN v2: Diverse Image Synthesis for Multiple Domains
• Generative adversarial networks for speech processing: A review
• GMM-UNIT: Unsupervised Multi-Domain and Multi-Modal Image-to-Image Translation via Attribute Gaussian Mixture Modeling
• Generative AI in Depth: A Survey of Recent Advances, Model Variants, and Real-World Applications
• Adaptive Hierarchical Dual Consistency for Semi-Supervised Left Atrium Segmentation on Cross-Domain Data
• Pose Augmentation: Class-agnostic Object Pose Transformation for Object Recognition
• Learning Fixed Points in Generative Adversarial Networks: From Image-to-Image Translation to Disease Detection and Localization
• Mix and match networks: cross-modal alignment for zero-pair image-to-image translation
• Recent Advances in Transfer Learning for Cross-Dataset Visual Recognition: A Problem-Oriented Perspective
• Neural Rerendering in the Wild
• A Review on Domain Adaption and Generative Adversarial Networks(GANs)
• Disentangled Cycle Consistency for Highly-realistic Virtual Try-On
• Show, Attend and Translate: Unpaired Multi-Domain Image-to-Image Translation with Visual Attention
• MixerGAN: An MLP-Based Architecture for Unpaired Image-to-Image Translation
• Contrastive Learning for Unpaired Image-to-Image Translation
• Unpaired Image-to-Image Translation using Adversarial Consistency Loss
• Creative synthesis of kinematic mechanisms
• Attribute Guided Unpaired Image-to-Image Translation with Semi-supervised Learning
• Universal Multi-Domain Translation via Diffusion Routers
• Generative Adversarial Networks Applied for Privacy Preservation in Biometric-Based Authentication and Identification
• Learning to see across Domains and Modalities
• Video-to-Video Translation for Visual Speech Synthesis
• Style Example-Guided Text Generation using Generative Adversarial Transformers
• ReshapeGAN: Object Reshaping by Providing A Single Reference Image
• MI²GAN: Generative Adversarial Network for Medical Image Domain Adaptation using Mutual Information Constraint
• DSRGAN: Explicitly Learning Disentangled Representation of Underlying Structure and Rendering for Image Generation without Tuple Supervision
• k-Same-Siamese-GAN: k-Same Algorithm with Generative Adversarial Network for Facial Image De-identification with Hyperparameter Tuning and Mixed Precision Training
• DA-Font: Few-Shot Font Generation via Dual-Attention Hybrid Integration
• RaceGAN: A Framework for Preserving Individuality while Converting Racial Information for Image-to-Image Translation
• CycleGAN-VC2: Improved CycleGAN-based Non-parallel Voice Conversion
• TSIT: A Simple and Versatile Framework for Image-to-Image Translation
• Ear2Face: Deep Biometric Modality Mapping
• Label-Noise Robust Generative Adversarial Networks
• Keep it Simple: Image Statistics Matching for Domain Adaptation
• LatentVoiceGrad: Nonparallel Voice Conversion with Latent Diffusion/Flow-Matching Models
• Unsupervised Enhancement of Real-World Depth Images Using Tri-Cycle GAN
• Semi-supervised Feature-Level Attribute Manipulation for Fashion Image Retrieval
• Image-to-Image Translation: Methods and Applications
• Moment Matching for Multi-Source Domain Adaptation
• Independent Encoder for Deep Hierarchical Unsupervised Image-to-Image Translation
• TRUST: Token-dRiven Ultrasound Style Transfer for Cross-Device Adaptation
• Generative Feature Replay For Class-Incremental Learning
• FCR: Investigating Generative AI models for Forensic Craniofacial Reconstruction
• Fine-Tuning StyleGAN2 For Cartoon Face Generation
• An Efficient Solution for Breast Tumor Segmentation and Classification in Ultrasound Images Using Deep Adversarial Learning
• Emotion Generation and Recognition: A StarGAN Approach
• Learning Layout and Style Reconfigurable GANs for Controllable Image Synthesis
• Distributed Generative Adversarial Net
• Towards Shape Biased Unsupervised Representation Learning for Domain Generalization
• CycleDiff: Cycle Diffusion Models for Unpaired Image-to-image Translation
• MU-GAN: Facial Attribute Editing based on Multi-attention Mechanism
• Cascade EF-GAN: Progressive Facial Expression Editing with Local Focuses
• StyDeco: Unsupervised Style Transfer with Distilling Priors and Semantic Decoupling
• Cross-Domain Cascaded Deep Feature Translation
• Unsupervised Multi-Domain Multimodal Image-to-Image Translation with Explicit Domain-Constrained Disentanglement
• VoiceGrad: Non-Parallel Any-to-Many Voice Conversion with Annealed Langevin Dynamics
• Network Architecture Search for Domain Adaptation
• Domain2Vec: Domain Embedding for Unsupervised Domain Adaptation
• Low-light Image Enhancement Algorithm Based on Retinex and Generative Adversarial Network
• Image-to-Image Translation of Synthetic Samples for Rare Classes
• Expression Conditional GAN for Facial Expression-to-Expression Translation
• SPIGAN: Privileged Adversarial Learning from Simulation
• Risk Bounds for Unsupervised Cross-Domain Mapping with IPMs
• Data-Driven Approach to Encoding and Decoding 3-D Crystal Structures
• Ancient Painting to Natural Image: A New Solution for Painting Processing
• Langevin Cooling for Domain Translation
• Unaligned Image-to-Image Translation by Learning to Reweight
• Deepfake Detection Scheme Based on Vision Transformer and Distillation
• Integrating Categorical Semantics into Unsupervised Domain Translation
• Dual Generator Generative Adversarial Networks for Multi-Domain Image-to-Image Translation
• CycleVAR: Repurposing Autoregressive Model for Unsupervised One-Step Image Translation
• One-to-one Mapping for Unpaired Image-to-image Translation
• Advancing Facial Stylization through Semantic Preservation Constraint and Pseudo-Paired Supervision
• Domain Adaptation with Morphologic Segmentation
• Style Transfer: A Decade Survey
• Diversify and Match: A Domain Adaptive Representation Learning Paradigm for Object Detection
• Unsupervised Many-to-Many Image-to-Image Translation Across Multiple Domains
• MangaGAN: Unpaired Photo-to-Manga Translation Based on The Methodology of Manga Drawing
• U-GAT-IT: Unsupervised Generative Attentional Networks with Adaptive Layer-Instance Normalization for Image-to-Image Translation
• Cross Domain Knowledge Transfer for Unsupervised Vehicle Re-identification
• One-Shot Image-to-Image Translation via Part-Global Learning with a Multi-adversarial Framework
• InstancePin: Instance-Addressable Layout-to-Image Diffusion via Coordinate Pinning
• MirrorNet: A Deep Bayesian Approach to Reflective 2D Pose Estimation from Human Images
• Medical Image Generation using Generative Adversarial Networks
• Sparse Pose Trajectory Completion
• Generative Adversarial Networks in Computer Vision: A Survey and Taxonomy
• Reusing Discriminators for Encoding: Towards Unsupervised Image-to-Image Translation
• GANHopper: Multi-Hop GAN for Unsupervised Image-to-Image Translation
• Crossing-Domain Generative Adversarial Networks for Unsupervised Multi-Domain Image-to-Image Translation
• On Deep Domain Adaptation: Some Theoretical Understandings
• Multi-Target Domain Adaptation via Unsupervised Domain Classification for Weather Invariant Object Detection
• When Autonomous Systems Meet Accuracy and Transferability through AI: A Survey
• TransferI2I: Transfer Learning for Image-to-Image Translation from Small Datasets
• DCDLearn: Multi-order Deep Cross-distance Learning for Vehicle Re-Identification
• MaskCycleGAN-VC: Learning Non-parallel Voice Conversion with Filling in Frames
• Learn to synthesize and synthesize to learn
• Nonparallel Voice Conversion with Augmented Classifier Star Generative Adversarial Networks
• Face-to-Music Translation Using a Distance-Preserving Generative Adversarial Network with an Auxiliary Discriminator
• Handloom Design Generation Using Generative Networks
• Multi-Domain Image Completion for Random Missing Input Data
• Cycle Consistent Probability Divergences Across Different Spaces
• Bayesian Cycle-Consistent Generative Adversarial Networks via Marginalizing Latent Sampling
• Federated Adversarial Domain Adaptation
• Similarity-preserving Image-image Domain Adaptation for Person Re-identification
• Point Cloud Completion by Learning Shape Priors
• Visual Diver Face Recognition for Underwater Human-Robot Interaction
• SingleGAN: Image-to-Image Translation by a Single-Generator Network using Multiple Generative Adversarial Learning
• Neural Wireframe Renderer: Learning Wireframe to Image Translations
• Potential Flow Generator with L₂ Optimal Transport Regularity for Generative Models
• The Spatially-Correlative Loss for Various Image Translation Tasks
• State of the Art on Neural Rendering
• Learning a Domain-Agnostic Visual Representation for Autonomous Driving via Contrastive Loss
• Unpaired Image Enhancement Featuring Reinforcement-Learning-Controlled Image Editing Software
• A Flexible Selection Scheme for Minimum-Effort Transfer Learning
• Adversarial-based neural networks for affect estimations in the wild
• Dual Adversarial Inference for Text-to-Image Synthesis
• Generative Adversarial Network with Multi-Branch Discriminator for Cross-Species Image-to-Image Translation
• Iconify: Converting Photographs into Icons
• Circular Image Deturbulence using Quasi-conformal Geometry
• Crossmodal Voice Conversion
• DeepKeyGen: A Deep Learning-based Stream Cipher Generator for Medical Image Encryption and Decryption
• Fake Generated Painting Detection Via Frequency Analysis
• Contextual Colorization and Denoising for Low-Light Ultra High Resolution Sequences

Discussions

• Learning to Discover Cross-Domain Relations with GANs [hn, 3 points, 0 comments]
• Learning to Discover Cross-Domain Relations with GANs [hn, 2 points, 0 comments]

Related

• Resembled Generative Adversarial Networks: Two Domains with Similar Attributes
• CoordGAN: Self-Supervised Dense Correspondences Emerge from GANs
• Generate To Adapt: Aligning Domains using Generative Adversarial Networks
• Semantically Decomposing the Latent Spaces of Generative Adversarial Networks
• Exploring Explicit Domain Supervision for Latent Space Disentanglement in Unpaired Image-to-Image Translation