Augmented CycleGAN: Learning Many-to-Many Mappings from Unpaired Data

Abstract

Learning inter-domain mappings from unpaired data can improve performance in structured prediction tasks, such as image segmentation, by reducing the need for paired data. CycleGAN was recently proposed for this problem, but critically assumes the underlying inter-domain mapping is approximately deterministic and one-to-one. This assumption renders the model ineffective for tasks requiring flexible, many-to-many mappings. We propose a new model, called Augmented CycleGAN, which learns many-to-many mappings between domains. We examine Augmented CycleGAN qualitatively and quantitatively on several image datasets.

arXiv commentICML 2018

References in corpus (12)

• Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks
• Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks
• InfoGAN: Interpretable Representation Learning by Information Maximizing Generative Adversarial Nets
• Toward Multimodal Image-to-Image Translation
• Style Transfer from Non-Parallel Text by Cross-Alignment
• Unsupervised Image-to-Image Translation Networks
• Adversarial Autoencoders
• DualGAN: Unsupervised Dual Learning for Image-to-Image Translation
• Unsupervised Machine Translation Using Monolingual Corpora Only
• CycleGAN, a Master of Steganography
• Generative Visual Manipulation on the Natural Image Manifold
• On the Effects of Batch and Weight Normalization in Generative Adversarial Networks

Cited by in corpus (94)

• Multimodal Unsupervised Image-to-Image Translation
• A Review on Generative Adversarial Networks: Algorithms, Theory, and Applications
• Disentangled Representation Learning in Cardiac Image Analysis
• DRIT++: Diverse Image-to-Image Translation via Disentangled Representations
• Contrastive Learning for Unpaired Image-to-Image Translation
• Enhanced Balancing GAN: Minority-class Image Generation
• Learning Disentangled Representations in the Imaging Domain
• InstaGAN: Instance-aware Image-to-Image Translation
• StarGAN v2: Diverse Image Synthesis for Multiple Domains
• Image-to-image translation for cross-domain disentanglement
• Few-Shot Unsupervised Image-to-Image Translation
• MirrorGAN: Learning Text-to-image Generation by Redescription
• Unsupervised Object Segmentation by Redrawing
• Pseudo-healthy synthesis with pathology disentanglement and adversarial learning
• Yes, we GAN: Applying Adversarial Techniques for Autonomous Driving
• Model Patching: Closing the Subgroup Performance Gap with Data Augmentation
• Augmented Cyclic Adversarial Learning for Low Resource Domain Adaptation
• Exemplar Guided Unsupervised Image-to-Image Translation with Semantic Consistency
• Generative Adversarial Networks for Image and Video Synthesis: Algorithms and Applications
• On the Evaluation of Conditional GANs
• Conditional Generation of Medical Images via Disentangled Adversarial Inference
• Self-Ensembling with GAN-based Data Augmentation for Domain Adaptation in Semantic Segmentation
• Using latent space regression to analyze and leverage compositionality in GANs
• CrystalGAN: Learning to Discover Crystallographic Structures with Generative Adversarial Networks
• Image-to-Image Translation: Methods and Applications
• Geometry-Consistent Generative Adversarial Networks for One-Sided Unsupervised Domain Mapping
• GMM-UNIT: Unsupervised Multi-Domain and Multi-Modal Image-to-Image Translation via Attribute Gaussian Mixture Modeling
• MW-GAN: Multi-Warping GAN for Caricature Generation with Multi-Style Geometric Exaggeration
• SEIGAN: Towards Compositional Image Generation by Simultaneously Learning to Segment, Enhance, and Inpaint
• Unsupervised Domain-Specific Deblurring via Disentangled Representations
• Domain Adaptive Relational Reasoning for 3D Multi-Organ Segmentation
• RL-CycleGAN: Reinforcement Learning Aware Simulation-To-Real
• Multi-Domain Translation by Learning Uncoupled Autoencoders
• Unsupervised Traffic Scene Generation with Synthetic 3D Scene Graphs
• Scalable Unbalanced Optimal Transport using Generative Adversarial Networks
• Optimal Unsupervised Domain Translation
• MISO: Mutual Information Loss with Stochastic Style Representations for Multimodal Image-to-Image Translation
• Cross-Domain Cascaded Deep Feature Translation
• Hypercomplex Image-to-Image Translation
• Recent Advances in 3D Object and Hand Pose Estimation
• LADN: Local Adversarial Disentangling Network for Facial Makeup and De-Makeup
• Neural Rerendering in the Wild
• Galaxy Image Translation with Semi-supervised Noise-reconstructed Generative Adversarial Networks
• CounteRGAN: Generating Realistic Counterfactuals with Residual Generative Adversarial Nets
• Emotional Voice Conversion: Theory, Databases and ESD
• Learning from Multi-domain Artistic Images for Arbitrary Style Transfer
• Open-Edit: Open-Domain Image Manipulation with Open-Vocabulary Instructions
• Domain-Scalable Unpaired Image Translation via Latent Space Anchoring
• Model-Based Domain Generalization
• AD-GAN: End-to-end Unsupervised Nuclei Segmentation with Aligned Disentangling Training
• Unpaired Pose Guided Human Image Generation
• SDIT: Scalable and Diverse Cross-domain Image Translation
• Dual Adversarial Inference for Text-to-Image Synthesis
• Label-Noise Robust Multi-Domain Image-to-Image Translation
• A Novel BiLevel Paradigm for Image-to-Image Translation
• PMC-GANs: Generating Multi-Scale High-Quality Pedestrian with Multimodal Cascaded GANs
• Diverse Semantic Image Synthesis via Probability Distribution Modeling
• Model-Based Robust Deep Learning: Generalizing to Natural, Out-of-Distribution Data
• Unpaired Image-to-Image Translation using Adversarial Consistency Loss
• ReLGAN: Generalization of Consistency for GAN with Disjoint Constraints and Relative Learning of Generative Processes for Multiple Transformation Learning
• Compositional Deep Learning
• Sketch-to-Art: Synthesizing Stylized Art Images From Sketches
• Joint Learning of Generative Translator and Classifier for Visually Similar Classes
• Controlling biases and diversity in diverse image-to-image translation
• Training Generative Adversarial Networks from Incomplete Observations using Factorised Discriminators
• SDA-GAN: Unsupervised Image Translation Using Spectral Domain Attention-Guided Generative Adversarial Network
• Unsupervised Vehicle Re-identification with Progressive Adaptation
• Conditional Denoising of Remote Sensing Imagery Using Cycle-Consistent Deep Generative Models
• Multimodal Image-to-Image Translation via Mutual Information Estimation and Maximization
• Visual-based Autonomous Driving Deployment from a Stochastic and Uncertainty-aware Perspective
• Augmented Cyclic Consistency Regularization for Unpaired Image-to-Image Translation
• DLOW: Domain Flow for Adaptation and Generalization
• Unsupervised Multimodal Video-to-Video Translation via Self-Supervised Learning
• Integrating Categorical Semantics into Unsupervised Domain Translation
• Generative Design of Hardware-aware DNNs
• Towards Learning a Self-inverse Network for Bidirectional Image-to-image Translation
• One-to-one Mapping for Unpaired Image-to-image Translation
• Bayesian Cycle-Consistent Generative Adversarial Networks via Marginalizing Latent Sampling
• Fashion-AttGAN: Attribute-Aware Fashion Editing with Multi-Objective GAN
• Unsupervised Adversarial Image Inpainting
• CoPE: Conditional image generation using Polynomial Expansions
• Unsupervised Many-to-Many Image-to-Image Translation Across Multiple Domains
• Generative Creativity: Adversarial Learning for Bionic Design
• A Novel Framework for Image-to-image Translation and Image Compression
• Neuronal Learning Analysis using Cycle-Consistent Adversarial Networks
• Semantic Example Guided Image-to-Image Translation
• Disentangled Unsupervised Image Translation via Restricted Information Flow
• Unaligned Image-to-Image Translation by Learning to Reweight
• Multimodal Image Synthesis with Conditional Implicit Maximum Likelihood Estimation
• An Unpaired Shape Transforming Method for Image Translation and Cross-Domain Retrieval
• Image-to-image Mapping with Many Domains by Sparse Attribute Transfer
• TransferI2I: Transfer Learning for Image-to-Image Translation from Small Datasets
• Domain-Agnostic Learning with Anatomy-Consistent Embedding for Cross-Modality Liver Segmentation
• Unpaired Deep Learning for Accelerated MRI using Optimal Transport Driven CycleGAN