Feature Re-Representation and Reliable Pseudo Label Retraining for Cross-Domain Semantic Segmentation
IEEE Transactions on Pattern Analysis and Machine Intelligence · 2024-03 · 46 (3) · 1682-1694
DOI: 10.1109/tpami.2022.3154933 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Transactions on Pattern Analysis and Machine Intelligence, 2024-03. Verfasst von Jing Li und 5 weiteren Autor:innen. Die Arbeit wurde bisher 13-mal zitiert (Crossref) und nennt 67 Literaturangaben. Gefördert von: National Key R&D Program of China, National Natural Science Foundation of China, Science and Technology Commission of Shanghai Municipality.
Abstract (Originalsprache)
This paper presents a novel unsupervised domain adaptation method for semantic segmentation. We argue that a good representation of the target-domain data should keep both the knowledge from the source domain and the target-domain-specific information. To obtain the knowledge from the source domain, we first learn a set of bases to characterize the feature distribution of the source domain, then features from both the source and the target domain are re-represented as a weighted summation of the source bases. A discriminator is additionally introduced to make the re-representation responsibilities of both domain features under the same bases indistinguishable. In this way, the domain gap between the source re-representation and target re-representation is minimized, and the re-represented target domain features contain the source domain information. Then we combine the feature re-representation with the original domain-specific feature together for subsequent pixel-wise classification. To further make the re-represented target features semantically meaningful, a Reliable Pseudo Label Retraining (RPLR) strategy is proposed, which utilizes the consistency of the prediction by the networks trained with multi-view source images to select the clean pseudo labels on unlabeled target images for re-training. Extensive experiments demonstrate the competitive performance of our approach for unsupervised domain adaptation on the semantic segmentation benchmarks.
Förderung
- National Key R&D Program of China
- National Natural Science Foundation of China
- Science and Technology Commission of Shanghai Municipality
- Shanghai Education Development Foundation and Shanghai Municipal Education Commission
- New Generation of AI
Zitieren
APA
Li, J., Zhou, K., Qian, S., Li, W., Duan, L., & Gao, S. (2024). Feature Re-Representation and Reliable Pseudo Label Retraining for Cross-Domain Semantic Segmentation. IEEE Transactions on Pattern Analysis and Machine Intelligence, 46(3), 1682-1694. https://doi.org/10.1109/tpami.2022.3154933
Vancouver
Li J, Zhou K, Qian S, Li W, Duan L, Gao S. Feature Re-Representation and Reliable Pseudo Label Retraining for Cross-Domain Semantic Segmentation. IEEE Transactions on Pattern Analysis and Machine Intelligence. 2024;46(3):1682-1694. doi:10.1109/tpami.2022.3154933
BibTeX
@article{Li2024,
author = {Li, Jing and Zhou, Kang and Qian, Shenhan and Li, Wen and Duan, Lixin and Gao, Shenghua},
title = {Feature Re-Representation and Reliable Pseudo Label Retraining for Cross-Domain Semantic Segmentation},
journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence},
year = {2024},
volume = {46},
number = {3},
pages = {1682--1694},
doi = {10.1109/tpami.2022.3154933}
}
EUVIMED