Semantic-Aware Multi-Scale Contrastive Learning for UAV Tiny Object Tracking
International Journal on Semantic Web and Information Systems · 08.06.2026 · 22 (1) · 1-24
DOI: 10.4018/ijswis.411908 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal on Semantic Web and Information Systems, erschienen am 08.06.2026. Verfasst von Xuehua Tao und 1 weiteren Autor:innen. Die Arbeit nennt 42 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 3.0).
Abstract (Originalsprache)
Unmanned aerial vehicle-based visual systems are essential for tasks such as reconnaissance, disaster monitoring, and traffic analysis; however, tracking remains challenging due to tiny targets, background clutter, and interference from similar objects. The authors propose multi-scale contrastive discrimination tracking, a transformer-based framework for tiny object tracking that integrates three modules: multi-scale detail enhancement, contrastive discrimination attention model for suppressing background interference via contrastive discrimination, and instance decoupling enhancement module for reducing feature coupling and identity drift among similar instances. Experiments on DTB70 and VisDrone2018 show that multi-scale contrastive discrimination tracking achieves 87.9% precision and a 67.4% success rate, surpassing state-of-the-art methods by 2.3% and 2.4%, respectively, while running at 185 frames per second on GPU with only 4.2 GMac and 11.2M parameters, demonstrating strong potential for resource-constrained unmanned arial vehicle deployment.
Zitieren
APA
Tao, X., & Sun, J. (2026). Semantic-Aware Multi-Scale Contrastive Learning for UAV Tiny Object Tracking. International Journal on Semantic Web and Information Systems, 22(1), 1-24. https://doi.org/10.4018/ijswis.411908
Vancouver
Tao X, Sun J. Semantic-Aware Multi-Scale Contrastive Learning for UAV Tiny Object Tracking. International Journal on Semantic Web and Information Systems. 2026;22(1):1-24. doi:10.4018/ijswis.411908
BibTeX
@article{Tao2026,
author = {Tao, Xuehua and Sun, Jiwei},
title = {Semantic-Aware Multi-Scale Contrastive Learning for UAV Tiny Object Tracking},
journal = {International Journal on Semantic Web and Information Systems},
year = {2026},
volume = {22},
number = {1},
pages = {1--24},
doi = {10.4018/ijswis.411908}
}
EUVIMED