CyCADA: Cycle-Consistent Adversarial Domain Adaptation

Abstract

Domain adaptation is critical for success in new, unseen environments. Adversarial adaptation models applied in feature spaces discover domain invariant representations, but are difficult to visualize and sometimes fail to capture pixel-level and low-level domain shifts. Recent work has shown that generative adversarial networks combined with cycle-consistency constraints are surprisingly effective at mapping images between domains, even without the use of aligned image pairs. We propose a novel discriminatively-trained Cycle-Consistent Adversarial Domain Adaptation model. CyCADA adapts representations at both the pixel-level and feature-level, enforces cycle-consistency while leveraging a task loss, and does not require aligned pairs. Our model can be applied in a variety of visual recognition and prediction settings. We show new state-of-the-art results across multiple adaptation tasks, including digit classification and semantic segmentation of road scenes demonstrating transfer from synthetic to real world domains.

References in corpus (19)

• Distilling the Knowledge in a Neural Network
• Conditional Generative Adversarial Nets
• Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks
• Unsupervised Domain Adaptation by Backpropagation
• Deep Domain Confusion: Maximizing for Domain Invariance
• Deep Generative Image Models using a Laplacian Pyramid of Adversarial Networks
• The Cityscapes Dataset for Semantic Urban Scene Understanding
• Energy-based Generative Adversarial Network
• FCNs in the Wild: Pixel-level Adversarial and Constraint-based Adaptation
• Adversarial Feature Learning
• Adversarially Learned Inference
• Adversarial Discriminative Domain Adaptation
• Unsupervised Cross-Domain Image Generation
• Deep CORAL: Correlation Alignment for Deep Domain Adaptation
• Dilated Residual Networks
• Playing for Data: Ground Truth from Computer Games
• Generative Visual Manipulation on the Natural Image Manifold
• Learning to Generate Images of Outdoor Scenes from Attributes and Semantic Layouts
• Scribbler: Controlling Deep Image Synthesis with Sketch and Color

Cited by in corpus (78)

• Semantic Foggy Scene Understanding with Synthetic Data
• SynthSeg: Segmentation of brain MRI scans of any contrast and resolution without retraining
• Aligning Domain-specific Distribution and Classifier for Cross-domain Classification from Multiple Sources
• How Generative Adversarial Networks and Their Variants Work: An Overview
• A Comprehensive Survey on Test-Time Adaptation under Distribution Shifts
• Unsupervised Domain Adaptation through Self-Supervision
• On Learning Invariant Representation for Domain Adaptation
• Can we trust deep learning models diagnosis? The impact of domain shift in chest radiograph classification
• SuperAnimal pretrained pose estimation models for behavioral analysis
• Uncertainty-Aware Consistency Regularization for Cross-Domain Semantic Segmentation
• Generative Domain Adaptation for Face Anti-Spoofing
• Source-Relaxed Domain Adaptation for Image Segmentation
• Transfer Learning for Related Reinforcement Learning Tasks via Image-to-Image Translation
• Bridging the Gap between Events and Frames through Unsupervised Domain Adaptation
• MT-UDA: Towards Unsupervised Cross-modality Medical Image Segmentation with Limited Source Labels
• Unsupervised Domain Adaptation for Object Detection via Cross-Domain Semi-Supervised Learning
• Learning Texture Invariant Representation for Domain Adaptation of Semantic Segmentation
• Self-Training and Adversarial Background Regularization for Unsupervised Domain Adaptive One-Stage Object Detection
• Towards Discriminability and Diversity: Batch Nuclear-norm Maximization under Label Insufficient Situations
• Significance-aware Information Bottleneck for Domain Adaptive Semantic Segmentation
• Complete & Label: A Domain Adaptation Approach to Semantic Segmentation of LiDAR Point Clouds
• All about Structure: Adapting Structural Information across Domains for Boosting Semantic Segmentation
• Towards Shape Biased Unsupervised Representation Learning for Domain Generalization
• Lifting 2D Human Pose to 3D with Domain Adapted 3D Body Concept
• Regularizing Proxies with Multi-Adversarial Training for Unsupervised Domain-Adaptive Semantic Segmentation
• The Direction-Aware, Learnable, Additive Kernels and the Adversarial Network for Deep Floor Plan Recognition
• RL-CycleGAN: Reinforcement Learning Aware Simulation-To-Real
• Unsupervised Video-to-Video Translation
• Gradually Vanishing Bridge for Adversarial Domain Adaptation
• Multi-source Pseudo-label Learning of Semantic Segmentation for the Scene Recognition of Agricultural Mobile Robots
• ACE: Adapting to Changing Environments for Semantic Segmentation
• Domain Adaptation Through Task Distillation
• TAFE-Net: Task-Aware Feature Embeddings for Low Shot Learning
• Anomaly Detection with Domain Adaptation
• Domain Alignment with Triplets
• Contrastive Examples for Addressing the Tyranny of the Majority
• Semi-supervised Domain Adaptation based on Dual-level Domain Mixing for Semantic Segmentation
• Meta-hallucinator: Towards Few-Shot Cross-Modality Cardiac Image Segmentation
• ST3D: Self-training for Unsupervised Domain Adaptation on 3D Object Detection
• FixBi: Bridging Domain Spaces for Unsupervised Domain Adaptation
• SOSELETO: A Unified Approach to Transfer Learning and Training with Noisy Labels
• Domain-Aware Dynamic Networks
• RetinaGAN: An Object-aware Approach to Sim-to-Real Transfer
• Unsupervised Domain Transfer with Conditional Invertible Neural Networks
• Unsupervised Domain Adaptation for Vestibular Schwannoma and Cochlea Segmentation via Semi-supervised Learning and Label Fusion
• Using Simulated Data to Generate Images of Climate Change
• BiPC: Bidirectional Probability Calibration for Unsupervised Domain Adaption
• Unsupervised domain adaptation via coarse-to-fine feature alignment method using contrastive learning
• Delving into Robust Object Detection from Unmanned Aerial Vehicles: A Deep Nuisance Disentanglement Approach
• Union-set Multi-source Model Adaptation for Semantic Segmentation
• Context-Aware Domain Adaptation in Semantic Segmentation
• CORE-ReID V2: Advancing the Domain Adaptation for Object Re-Identification with Optimized Training and Ensemble Fusion
• TuneNet: One-Shot Residual Tuning for System Identification and Sim-to-Real Robot Task Transfer
• Learning Cross-Domain Correspondence for Control with Dynamics Cycle-Consistency
• Domain Adaptation and Image Classification via Deep Conditional Adaptation Network
• Unsupervised Domain Adaptation: from Simulation Engine to the RealWorld
• A Survey of Techniques All Classifiers Can Learn from Deep Networks: Models, Optimizations, and Regularization
• Invenio: Discovering Hidden Relationships Between Tasks/Domains Using Structured Meta Learning
• A Study into Echocardiography View Conversion
• Generalizable Model-agnostic Semantic Segmentation via Target-specific Normalization
• Adaptively-Accumulated Knowledge Transfer for Partial Domain Adaptation
• Multilevel Knowledge Transfer for Cross-Domain Object Detection
• Generation For Adaption: A GAN-Based Approach for 3D Domain Adaption with Point Cloud Data
• Symbiotic Adversarial Learning for Attribute-based Person Search
• Large-Scale Optimal Transport via Adversarial Training with Cycle-Consistency
• Improved Surrogates in Inertial Confinement Fusion with Manifold and Cycle Consistencies
• Fast Computational Ghost Imaging using Unpaired Deep Learning and a Constrained Generative Adversarial Network
• Exploiting Inter-pixel Correlations in Unsupervised Domain Adaptation for Semantic Segmentation
• Memory-guided Unsupervised Image-to-image Translation
• Dual Domain-Adversarial Learning for Audio-Visual Saliency Prediction
• LRS-DAG: Low Resource Supervised Domain Adaptation with Generalization Across Domains
• Enlisting 3D Crop Models and GANs for More Data Efficient and Generalizable Fruit Detection
• Adversarial Augmentation Training Makes Action Recognition Models More Robust to Realistic Video Distribution Shifts
• Learning to see across Domains and Modalities
• Detection and Segmentation of Custom Objects using High Distraction Photorealistic Synthetic Data
• Separating Content and Style for Unsupervised Image-to-Image Translation
• Reproducibility of "FDA: Fourier Domain Adaptation forSemantic Segmentation
• Synthesizing Photorealistic Images with Deep Generative Learning