EUVIMEDEUVIMED entdecken
Crossref

Efficient One-Shot Gesture Recognition for WiFi ISAC via Aug-Meta Learning

Jianwei Liu, Jiantao Yuan, Guanding Yu, Jinsong Han

IEEE Journal on Selected Areas in Communications · 2025-11 · 43 (11) · 3766-3781

DOI: 10.1109/jsac.2025.3584433 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Journal on Selected Areas in Communications, 2025-11. Verfasst von Jianwei Liu und 3 weiteren Autor:innen. Die Arbeit wurde bisher 5-mal zitiert (Crossref) und nennt 76 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

WiFi-based gesture recognition (WGR) has emerged as a promising technology due to its potential for integration with communication systems under the concept of integrated sensing and communication (ISAC). However, current WGR systems face two primary challenges: limited scalability for recognizing new gestures and poor compatibility with ISAC. These systems typically require extensive data collection and retraining for each new gesture and struggle to handle the dimensional variability of channel state information (CSI) caused by fluctuating data traffic in communication networks. To overcome these limitations, we introduce OneSense, a one-shot WGR system designed for seamless integration with communication systems. OneSense designs a data enrichment technique based on the law of signal propagation to generate virtual gestures. Based on enriched dataset, OneSense leverages an aug-meta learning (AML) framework to facilitate efficient and scalable FSL. OneSense also incorporates a data cropping strategy to enhance gesture feature prominence and a dynamic size-adaptive backbone model that ensures compatibility with CSI samples exhibiting dimensional inconsistencies. Experimental results show that OneSense achieves over 94% accuracy in one-shot gesture recognition. A case study further illustrates its effectiveness in ISAC contexts. Furthermore, our proposed AML framework reduces pre-training latency by more than 86% compared to conventional meta-learning approaches.

Zitieren

APA

Liu, J., Yuan, J., Yu, G., & Han, J. (2025). Efficient One-Shot Gesture Recognition for WiFi ISAC via Aug-Meta Learning. IEEE Journal on Selected Areas in Communications, 43(11), 3766-3781. https://doi.org/10.1109/jsac.2025.3584433

Vancouver

Liu J, Yuan J, Yu G, Han J. Efficient One-Shot Gesture Recognition for WiFi ISAC via Aug-Meta Learning. IEEE Journal on Selected Areas in Communications. 2025;43(11):3766-3781. doi:10.1109/jsac.2025.3584433

BibTeX
@article{Liu2025,
  author = {Liu, Jianwei and Yuan, Jiantao and Yu, Guanding and Han, Jinsong},
  title = {Efficient One-Shot Gesture Recognition for WiFi ISAC via Aug-Meta Learning},
  journal = {IEEE Journal on Selected Areas in Communications},
  year = {2025},
  volume = {43},
  number = {11},
  pages = {3766--3781},
  doi = {10.1109/jsac.2025.3584433}
}
Kostenlos starten →