EUVIMEDEUVIMED entdecken
Crossref

Bridge: Leveraging Vision Foundation Models for Efficient Cross-Domain Remote Sensing Segmentation

Bin Wang, Zeyu Chen, Zhi Li, Ling Lei, Yiguang Liu

IEEE Transactions on Geoscience and Remote Sensing · 2026 · 64 · 1-13

DOI: 10.1109/tgrs.2026.3673710 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Geoscience and Remote Sensing, 2026. Verfasst von Bin Wang und 4 weiteren Autor:innen. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 62 Literaturangaben. Gefördert von: NSFC, National Key Research and Development Program of China, Science and Technology Funding of Sichuan Province.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Adapting large-scale vision foundation models (VFMs), pretrained on natural images (NIs), to remote sensing (RS) downstream tasks such as domain adaptation semantic segmentation (DASS) through parameter-efficient fine-tuning (PEFT) is a promising yet challenging direction. However, due to the large domain gap between NI and RS images (RSIs), as well as the inherent intradomain variations in the RS DASS task, existing PEFT methods struggle to achieve satisfactory adaptation. To address these challenges, we propose Bridge, a framework with small trainable parameters specifically designed for RS DASS. Bridge introduces a set of learnable domain-invariant tokens (DITs) that capture transferable knowledge shared across NI and RSI domains. To complement this, a spatial prior module (SPM) is proposed to extract multiscale spatial priors that provide auxiliary guidance for DIT adaptation (DITA), which produces domain-specific tokens. Finally, domain-aware injection (DAI) injects the domain-specific tokens into intermediate VFM features, aligning them with RS distributions and facilitating effective DASS task adaptation. With significantly fewer trainable parameters, Bridge efficiently transfers NI-pretrained VFMs to the RS DASS task. Extensive experiments demonstrate that Bridge substantially outperforms state-of-the-art PEFT approaches. Remarkably, compared with the current SOTA, Bridge achieves a significant improvement in mIoU. The code will be released at Bridge.

Förderung

  • NSFC
  • National Key Research and Development Program of China
  • Science and Technology Funding of Sichuan Province

Zitieren

APA

Wang, B., Chen, Z., Li, Z., Lei, L., & Liu, Y. (2026). Bridge: Leveraging Vision Foundation Models for Efficient Cross-Domain Remote Sensing Segmentation. IEEE Transactions on Geoscience and Remote Sensing, 64, 1-13. https://doi.org/10.1109/tgrs.2026.3673710

Vancouver

Wang B, Chen Z, Li Z, Lei L, Liu Y. Bridge: Leveraging Vision Foundation Models for Efficient Cross-Domain Remote Sensing Segmentation. IEEE Transactions on Geoscience and Remote Sensing. 2026;64:1-13. doi:10.1109/tgrs.2026.3673710

BibTeX
@article{Wang2026,
  author = {Wang, Bin and Chen, Zeyu and Li, Zhi and Lei, Ling and Liu, Yiguang},
  title = {Bridge: Leveraging Vision Foundation Models for Efficient Cross-Domain Remote Sensing Segmentation},
  journal = {IEEE Transactions on Geoscience and Remote Sensing},
  year = {2026},
  volume = {64},
  pages = {1--13},
  doi = {10.1109/tgrs.2026.3673710}
}
Kostenlos starten →