EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Mask-Guided Teacher–Student Learning for Open-Vocabulary Object Detection in Remote Sensing Images

Shuojie Wang, Yu Song, Jiajun Xiang, Yanyan Chen, Ping Zhong, Ruigang Fu

Remote Sensing · 09.10.2025 · 17 (19) · 3385

DOI: 10.3390/rs17193385 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Remote Sensing, online erschienen am 09.10.2025. Verfasst von Shuojie Wang und 5 weiteren Autor:innen. Die Arbeit wurde bisher 4-mal zitiert (Crossref) und nennt 29 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Fundamental Research Foundation of National Key Laboratory of Automatic Target Recognition.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Open-vocabulary object detection in remote sensing aims to detect novel categories not seen during training, which is crucial for practical aerial image analysis applications. While some approaches accomplish this task through large-scale data construction, such methods incur substantial annotation and computational costs. In contrast, we focus on efficient utilization of limited datasets. However, existing methods such as CastDet struggle with inefficient data utilization and class imbalance issues in pseudo-label generation for novel categories. We propose an enhanced open-vocabulary detection framework that addresses these limitations through two key innovations. First, we introduce a selective masking strategy that enables direct utilization of partially annotated images by masking base category regions in teacher model inputs. This approach eliminates the need for strict data separation and significantly improves data efficiency. Second, we develop a dynamic frequency-based class weighting that automatically adjusts category weights based on real-time pseudo-label statistics to mitigate class imbalance issues. Our approach integrates these components into a student–teacher learning framework with RemoteCLIP for novel category classification. Comprehensive experiments demonstrate significant improvements on both datasets: on VisDroneZSD, we achieve 42.7% overall mAP and 41.4% harmonic mean, substantially outperforming existing methods. On DIOR dataset, our method achieves 63.7% overall mAP with 49.5% harmonic mean. Our framework achieves more balanced performance between base and novel categories, providing a practical and data-efficient solution for open-vocabulary aerial object detection.

Förderung

  • Fundamental Research Foundation of National Key Laboratory of Automatic Target Recognition

Zitieren

APA

Wang, S., Song, Y., Xiang, J., Chen, Y., Zhong, P., & Fu, R. (2025). Mask-Guided Teacher–Student Learning for Open-Vocabulary Object Detection in Remote Sensing Images. Remote Sensing, 17(19), 3385. https://doi.org/10.3390/rs17193385

Vancouver

Wang S, Song Y, Xiang J, Chen Y, Zhong P, Fu R. Mask-Guided Teacher–Student Learning for Open-Vocabulary Object Detection in Remote Sensing Images. Remote Sensing. 2025;17(19):3385. doi:10.3390/rs17193385

BibTeX
@article{Wang2025,
  author = {Wang, Shuojie and Song, Yu and Xiang, Jiajun and Chen, Yanyan and Zhong, Ping and Fu, Ruigang},
  title = {Mask-Guided Teacher–Student Learning for Open-Vocabulary Object Detection in Remote Sensing Images},
  journal = {Remote Sensing},
  year = {2025},
  volume = {17},
  number = {19},
  pages = {3385},
  doi = {10.3390/rs17193385}
}
Kostenlos starten →