EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

PLAID: Pyramid-Level Adaptive Imbalance-Aware Diversification for Source-Free Domain Adaptation in Power-Grid Temporal Action Detection

Lingwen Meng, Siwu Yu, Mingyong Xin, Xinshan Zhu

Electronics · 18.11.2026 · 15 (22) · 5377

DOI: 10.3390/electronics15225377 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Electronics, online erschienen am 18.11.2026. Verfasst von Lingwen Meng und 3 weiteren Autor:innen. Die Arbeit nennt 47 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Guizhou Power Grid Co., Ltd.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Temporal action detection (TAD) localizes safety-critical procedures in untrimmed power-grid operation videos. Cross-site changes in action duration and density create level-dependent feature collapse in temporal pyramids and noisy, incomplete pseudo-labels under source-free domain adaptation. We propose Pyramid-Level Adaptive Imbalance-Aware Diversification (PLAID), which augments Mean-Teacher self-training with pyramid-level representation regularization. Duration-Aware Sampling (DAS) first constructs token matrices at duration-matched pyramid levels. Cross-Level Borrowing (CLB) then supplies temporally aligned neighboring tokens to levels with limited pseudo-segment support, stabilizing the statistics used by subsequent constraints. Channel variance and decorrelation losses preserve feature diversity, while Rank-Gap Reweighting (RGR) scales these constraints according to each level’s effective-rank decline from its running maximum. Together, these label-free components reduce reliance on imperfect pseudo-labels without source videos or target annotations. On four controlled duration and density shifts in the Power Grid Operations benchmark, PLAID improves source-only ActionFormer by 2.5–6.0 mean average precision (mAP) points and outperforms Vanilla Mean-Teacher in every setting. With TemporalMaxer, it also surpasses both baselines at every evaluated temporal intersection-over-union (tIoU) threshold in two representative shifts. The results support pyramid-aware representation recovery as a practical strategy for adapting fixed-camera power-grid detectors across sites when operational tempo and action occupancy vary and source recordings cannot be shared.

Förderung

  • Guizhou Power Grid Co., Ltd

Zitieren

APA

Meng, L., Yu, S., Xin, M., & Zhu, X. (2026). PLAID: Pyramid-Level Adaptive Imbalance-Aware Diversification for Source-Free Domain Adaptation in Power-Grid Temporal Action Detection. Electronics, 15(22), 5377. https://doi.org/10.3390/electronics15225377

Vancouver

Meng L, Yu S, Xin M, Zhu X. PLAID: Pyramid-Level Adaptive Imbalance-Aware Diversification for Source-Free Domain Adaptation in Power-Grid Temporal Action Detection. Electronics. 2026;15(22):5377. doi:10.3390/electronics15225377

BibTeX
@article{Meng2026,
  author = {Meng, Lingwen and Yu, Siwu and Xin, Mingyong and Zhu, Xinshan},
  title = {PLAID: Pyramid-Level Adaptive Imbalance-Aware Diversification for Source-Free Domain Adaptation in Power-Grid Temporal Action Detection},
  journal = {Electronics},
  year = {2026},
  volume = {15},
  number = {22},
  pages = {5377},
  doi = {10.3390/electronics15225377}
}
Kostenlos starten →