Contrastive Adaptation Network for Unsupervised Domain Adaptation

Guoliang Kang, Lu Jiang, Yi Yang, Alexander G. Hauptmann

2019-01-04CVPR 2019 6Unsupervised Domain Adaptation Domain Adaptation

Abstract

Unsupervised Domain Adaptation (UDA) makes predictions for the target domain data while manual annotations are only available in the source domain. Previous methods minimize the domain discrepancy neglecting the class information, which may lead to misalignment and poor generalization performance. To address this issue, this paper proposes Contrastive Adaptation Network (CAN) optimizing a new metric which explicitly models the intra-class domain discrepancy and the inter-class domain discrepancy. We design an alternating update strategy for training CAN in an end-to-end manner. Experiments on two real-world benchmarks Office-31 and VisDA-2017 demonstrate that CAN performs favorably against the state-of-the-art methods and produces more discriminative features.

Results

Task	Dataset	Metric	Value	Model
Domain Adaptation	Office-31	Average Accuracy	90.6	Contrastive Adaptation Network
Domain Adaptation	VisDA2017	Accuracy	87.2	CAN

Related Papers

A Privacy-Preserving Semantic-Segmentation Method Using Domain-Adaptation Technique2025-07-17 Domain Borders Are There to Be Crossed With Federated Few-Shot Adaptation2025-07-14 An Offline Mobile Conversational Agent for Mental Health Support: Learning from Emotional Dialogues and Psychological Texts with Student-Centered Evaluation2025-07-11 The Bayesian Approach to Continual Learning: An Overview2025-07-11 Doodle Your Keypoints: Sketch-Based Few-Shot Keypoint Detection2025-07-10 YOLO-APD: Enhancing YOLOv8 for Robust Pedestrian Detection on Complex Road Geometries2025-07-07 CORE-ReID V2: Advancing the Domain Adaptation for Object Re-Identification with Optimized Training and Ensemble Fusion2025-07-04 Underwater Monocular Metric Depth Estimation: Real-World Benchmarks and Synthetic Fine-Tuning2025-07-02