EUVIMEDEUVIMED entdecken
Crossref

Enhancing the Two-Stream Framework for Efficient Visual Tracking

Chengao Zong, Xin Chen, Jie Zhao, Yang Liu, Huchuan Lu, Dong Wang

IEEE Transactions on Image Processing · 2025 · 34 · 5500-5512

DOI: 10.1109/tip.2025.3598934 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Image Processing, 2025. Verfasst von Chengao Zong und 5 weiteren Autor:innen. Die Arbeit wurde bisher 5-mal zitiert (Crossref) und nennt 67 Literaturangaben. Gefördert von: National Natural Science Foundation of China, Fundamental Scientific Research Funding of the Central Universities of China, Liaoning Provincial Science and Technology Joint Program Project.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Practical deployments, especially on resource-limited edge devices, necessitate high speed for visual object trackers. To meet this demand, we introduce a new efficient tracker with a Two-Stream architecture, named ToS. While the recent one-stream tracking framework, employing a unified backbone for simultaneous processing of both the template and search region, has demonstrated exceptional efficacy, we find the conventional two-stream tracking framework, which employs two separate backbones for the template and search region, offers inherent advantages. The two-stream tracking framework is more compatible with advanced lightweight backbones and can efficiently utilize benefits from large templates. We demonstrate that the two-stream setup can exceed the one-stream tracking model in both speed and accuracy through strategic designs. Our methodology rejuvenates the two-stream tracking paradigm with lightweight pre-trained backbones and the proposed three efficient strategies: 1) A feature-aggregation module that improves the representation capability of the backbone, 2) A channel-wise approach for feature fusion, presenting a more effective and lighter alternative to spatial concatenation techniques, and 3) An expanded template strategy to boost tracking accuracy with negligible additional computational cost. Extensive evaluations across multiple tracking benchmarks demonstrate that the proposed method sets a new state-of-the-art performance in efficient visual tracking.

Förderung

  • National Natural Science Foundation of China
  • Fundamental Scientific Research Funding of the Central Universities of China
  • Liaoning Provincial Science and Technology Joint Program Project

Zitieren

APA

Zong, C., Chen, X., Zhao, J., Liu, Y., Lu, H., & Wang, D. (2025). Enhancing the Two-Stream Framework for Efficient Visual Tracking. IEEE Transactions on Image Processing, 34, 5500-5512. https://doi.org/10.1109/tip.2025.3598934

Vancouver

Zong C, Chen X, Zhao J, Liu Y, Lu H, Wang D. Enhancing the Two-Stream Framework for Efficient Visual Tracking. IEEE Transactions on Image Processing. 2025;34:5500-5512. doi:10.1109/tip.2025.3598934

BibTeX
@article{Zong2025,
  author = {Zong, Chengao and Chen, Xin and Zhao, Jie and Liu, Yang and Lu, Huchuan and Wang, Dong},
  title = {Enhancing the Two-Stream Framework for Efficient Visual Tracking},
  journal = {IEEE Transactions on Image Processing},
  year = {2025},
  volume = {34},
  pages = {5500--5512},
  doi = {10.1109/tip.2025.3598934}
}
Kostenlos starten →