From source to target and back: symmetric bi-directional adaptive GAN

Abstract

The effectiveness of generative adversarial approaches in producing images according to a specific style or visual domain has recently opened new directions to solve the unsupervised domain adaptation problem. It has been shown that source labeled images can be modified to mimic target samples making it possible to train directly a classifier in the target domain, despite the original lack of annotated data. Inverse mappings from the target to the source domain have also been evaluated but only passing through adapted feature spaces, thus without new image generation. In this paper we propose to better exploit the potential of generative adversarial networks for adaptation by introducing a novel symmetric mapping among domains. We jointly optimize bi-directional image transformations combining them with target self-labeling. Moreover we define a new class consistency loss that aligns the generators in the two directions imposing to conserve the class identity of an image passing through both domain mappings. A detailed qualitative and quantitative analysis of the reconstructed images confirm the power of our approach. By integrating the two domain specific classifiers obtained with our bi-directional network we exceed previous state-of-the-art unsupervised adaptation results on four different benchmark datasets.

References in corpus (11)

• Adam: A Method for Stochastic Optimization
• Conditional Generative Adversarial Nets
• Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks
• Conditional Image Synthesis With Auxiliary Classifier GANs
• Improved Techniques for Training GANs
• Domain Separation Networks
• Adversarial Discriminative Domain Adaptation
• Least Squares Generative Adversarial Networks
• Return of Frustratingly Easy Domain Adaptation
• Associative Domain Adaptation
• Domain Adaptation of Majority Votes via Perturbed Variation-based Label Transfer

Cited by in corpus (6)

• Self-ensembling for visual domain adaptation
• Synergistic Image and Feature Adaptation: Towards Cross-Modality Domain Adaptation for Medical Image Segmentation
• Cluster Alignment with a Teacher for Unsupervised Domain Adaptation
• Deep Adversarial Attention Alignment for Unsupervised Domain Adaptation: the Benefit of Target Expectation Maximization
• Joint Learning of Neural Networks via Iterative Reweighted Least Squares
• Parameter Reference Loss for Unsupervised Domain Adaptation