EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

SURA-SORT: Saving Unmatched Tracks With Re-ID and BUSCA-Integrated OC-SORT for Object Tracking

Thi-Nhung Le, Ba-Chau Tran, Duy-Tung Nguyen, van-Hieu Nguyen, Thi-Lan Le

IEEE Open Journal of Intelligent Transportation Systems · 2026 · 7 · 1687-1711

DOI: 10.1109/ojits.2026.3708477 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Open Journal of Intelligent Transportation Systems, 2026. Verfasst von Thi-Nhung Le und 4 weiteren Autor:innen. Die Arbeit nennt 55 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Hanoi University of Science and Technology, Vietnam.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Multi-object tracking (MOT) in videos is a fundamental task in computer vision and has been widely applied across various domains. The tracking-by-detection (TbD) method has become a popular approach due to its flexibility and scalability; however, its performance heavily depends on the quality of detections and the robustness of the association strategy. The OC-SORT object tracking algorithm, built on an observation-centric idea to update and recover trajectories during temporary track losses, is considered robust in handling occlusions and nonlinear object motion. Nevertheless, this method still faces challenges such as long-term occlusions, significant shape variations, and multiple objects with similar appearances. In this paper, we propose a novel method, SURA-SORT (Saving Unmatched Tracks with Re-ID and BUSCA-integrated OC-SORT for Object Tracking), which integrates OC-SORT with a BUSCA module for attention-driven continuous tracking and a Re-ID module for appearance-based re-identification (Re-ID), effectively addressing these challenges. Experimental results on two traffic surveillance datasets, VNTraffic and AICC22-Custom, demonstrate high tracking accuracy, achieving MOTA scores of 93.7% and 75.4%, respectively. These results highlight the strong potential of the proposed method for applications in ITS. Furthermore, extensive experiments on videos recorded at honeybee hive entrances indicate that the proposed method can be further improved for tracking this insect species with highly complex motion behaviors.

Förderung

  • Hanoi University of Science and Technology, Vietnam

Zitieren

APA

Le, T., Tran, B., Nguyen, D., Nguyen, V., & Le, T. (2026). SURA-SORT: Saving Unmatched Tracks With Re-ID and BUSCA-Integrated OC-SORT for Object Tracking. IEEE Open Journal of Intelligent Transportation Systems, 7, 1687-1711. https://doi.org/10.1109/ojits.2026.3708477

Vancouver

Le T, Tran B, Nguyen D, Nguyen V, Le T. SURA-SORT: Saving Unmatched Tracks With Re-ID and BUSCA-Integrated OC-SORT for Object Tracking. IEEE Open Journal of Intelligent Transportation Systems. 2026;7:1687-1711. doi:10.1109/ojits.2026.3708477

BibTeX
@article{Le2026,
  author = {Le, Thi-Nhung and Tran, Ba-Chau and Nguyen, Duy-Tung and Nguyen, van-Hieu and Le, Thi-Lan},
  title = {SURA-SORT: Saving Unmatched Tracks With Re-ID and BUSCA-Integrated OC-SORT for Object Tracking},
  journal = {IEEE Open Journal of Intelligent Transportation Systems},
  year = {2026},
  volume = {7},
  pages = {1687--1711},
  doi = {10.1109/ojits.2026.3708477}
}
Kostenlos starten →