All‐in‐One Video Restoration With Adaptive Degradation Perception and Global–Local Synergy
International Journal of Intelligent Systems · 2026-01 · 2026 (1)
DOI: 10.1155/int/8330880 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Intelligent Systems, online erschienen am 09.11.2026. Verfasst von Deqian Mao und 4 weiteren Autor:innen. Die Arbeit nennt 45 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Natural Science Foundation of Shandong Province, National Natural Science Foundation of China.
Abstract (Originalsprache)
All‐in‐one video restoration aims to restore degraded videos with different degradation types using a unified algorithm or model. However, existing video restoration methods have limited perception of video degradation types, making them unable to flexibly adapt to diverse degradation scenarios. Meanwhile, inaccurate calculation of cross‐temporal features between video frames affects the accuracy of video restoration. To address these issues, we propose an all‐in‐one video restoration method (DPGA‐VR) via adaptive degradation perception and global–local synergy, which fully mines the global–local representations between video frames and adaptively perceives different degradation types in degraded videos, breaking the traditional paradigm of “single model processing single degradation.” Specifically, a spatiotemporal feature encoder based on a local transformer accurately mines and extracts the global–local spatiotemporal features of the degraded video, achieving cross‐temporal perception between video frames. The adaptive degradation perceptron uses different “experts” to flexibly handle videos with different degradation types. In addition, the global–local synergy reconstruction can maximize the global–local consistency of the restored video. Extensive experiments on video deraining and dehazing tasks demonstrate that DPGA‐VR exhibits good all‐in‐one video restoration performance.
Förderung
- Natural Science Foundation of Shandong Province
- National Natural Science Foundation of China
Zitieren
APA
Mao, D., Gao, S., Jiao, X., Yao, X., & Zhou, Y. (2026). All‐in‐One Video Restoration With Adaptive Degradation Perception and Global–Local Synergy. International Journal of Intelligent Systems, 2026(1). https://doi.org/10.1155/int/8330880
Vancouver
Mao D, Gao S, Jiao X, Yao X, Zhou Y. All‐in‐One Video Restoration With Adaptive Degradation Perception and Global–Local Synergy. International Journal of Intelligent Systems. 2026;2026(1). doi:10.1155/int/8330880
BibTeX
@article{Mao2026,
author = {Mao, Deqian and Gao, Shanshan and Jiao, Xinchen and Yao, Xunxiang and Zhou, Yuanfeng},
title = {All‐in‐One Video Restoration With Adaptive Degradation Perception and Global–Local Synergy},
journal = {International Journal of Intelligent Systems},
year = {2026},
volume = {2026},
number = {1},
doi = {10.1155/int/8330880}
}
EUVIMED