We present a novel unsupervised domain adaptation method for semantic segmentation that generalizes a model trained with source images and corresponding ground-truth labels to a target domain. A key to domain adaptive semantic segmentation is to learn domain-invariant and discriminative features without target ground-truth labels. To this end, we propose a bi-directional pixel-prototype contrastive learning framework that minimizes intra-class variations of features for the same object class, while maximizing inter-class variations for different ones, regardless of domains. Specifically, our framework aligns pixel-level features and a prototype of the same object class in target and source images (i.e., positive pairs), respectively, sets them apart for different classes (i.e., negative pairs), and performs the alignment and separation processes toward the other direction with pixel-level features in the source image and a prototype in the target image. The cross-domain matching encourages domain-invariant feature representations, while the bidirectional pixel-prototype correspondences aggregate features for the same object class, providing discriminative features. To establish training pairs for contrastive learning, we propose to generate dynamic pseudo labels of target images using a non-parametric label transfer, that is, pixel-prototype correspondences across different domains. We also present a calibration method compensating class-wise domain biases of prototypes gradually during training. Experimental results on standard benchmarks including GTA5 → Cityscapes and SYNTHIA → Cityscapes demonstrate the effectiveness of our framework.
|Title of host publication||Computer Vision – ECCV 2022 - 17th European Conference, Proceedings|
|Editors||Shai Avidan, Gabriel Brostow, Moustapha Cissé, Giovanni Maria Farinella, Tal Hassner|
|Publisher||Springer Science and Business Media Deutschland GmbH|
|Number of pages||18|
|Publication status||Published - 2022|
|Event||17th European Conference on Computer Vision, ECCV 2022 - Tel Aviv, Israel|
Duration: 2022 Oct 23 → 2022 Oct 27
|Name||Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)|
|Conference||17th European Conference on Computer Vision, ECCV 2022|
|Period||22/10/23 → 22/10/27|
Bibliographical notePublisher Copyright:
© 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.
All Science Journal Classification (ASJC) codes
- Theoretical Computer Science
- Computer Science(all)