EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

End-to-End Online Video Stitching and Stabilization Method Based on Unsupervised Deep Learning

Pengyuan Wang, Pinle Qin, Rui Chai, Jianchao Zeng, Pengcheng Zhao, Zuojun Chen, Bingjie Han

Applied Sciences · 26.05.2025 · 15 (11) · 5987

DOI: 10.3390/app15115987 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Applied Sciences, online erschienen am 26.05.2025. Verfasst von Pengyuan Wang und 6 weiteren Autor:innen. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 19 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Major Science and Technology Project of Shanxi Province, “Unveiling the Leader” Project: Key Technologies and Demonstration for the Construction of a Green Airport and Intelligent Management and Control Platform.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The limited field of view, cumulative inter-frame jitter, and dynamic parallax interference in handheld video stitching often lead to misalignment and distortion. In this paper, we propose an end-to-end, unsupervised deep-learning framework that jointly performs real-time video stabilization and stitching. First, collaborative optimization architecture allows the stabilization and stitching modules to share parameters and propagate errors through a fully differentiable network, ensuring consistent image alignment. Second, a Markov trajectory smoothing strategy in relative coordinates models inter-frame motion as incremental relationships, effectively reducing cumulative errors. Third, a dynamic attention mask generates spatiotemporal weight maps based on foreground motion prediction, suppressing misalignment caused by dynamic objects. Experimental evaluation on diverse handheld sequences shows that our method achieves higher stitching quality, lower geometric distortion rates, and improved video stability compared to state-of-the-art baselines, while maintaining real-time processing capabilities. Ablation studies validate that relative trajectory modeling substantially mitigates long-term jitter and that the dynamic attention mask enhances stitching accuracy in dynamic scenes. These results demonstrate that the proposed framework provides a robust solution for high-quality, real-time handheld video stitching.

Förderung

  • Major Science and Technology Project of Shanxi Province
  • “Unveiling the Leader” Project: Key Technologies and Demonstration for the Construction of a Green Airport and Intelligent Management and Control Platform

Zitieren

APA

Wang, P., Qin, P., Chai, R., Zeng, J., Zhao, P., Chen, Z., & Han, B. (2025). End-to-End Online Video Stitching and Stabilization Method Based on Unsupervised Deep Learning. Applied Sciences, 15(11), 5987. https://doi.org/10.3390/app15115987

Vancouver

Wang P, Qin P, Chai R, Zeng J, Zhao P, Chen Z, et al. End-to-End Online Video Stitching and Stabilization Method Based on Unsupervised Deep Learning. Applied Sciences. 2025;15(11):5987. doi:10.3390/app15115987

BibTeX
@article{Wang2025,
  author = {Wang, Pengyuan and Qin, Pinle and Chai, Rui and Zeng, Jianchao and Zhao, Pengcheng and Chen, Zuojun and Han, Bingjie},
  title = {End-to-End Online Video Stitching and Stabilization Method Based on Unsupervised Deep Learning},
  journal = {Applied Sciences},
  year = {2025},
  volume = {15},
  number = {11},
  pages = {5987},
  doi = {10.3390/app15115987}
}
Kostenlos starten →