EUVIMEDEUVIMED entdecken
Crossref

Leveraging Multi-Image Representations for Improved Time-Series Knowledge Distillation on Wearable Sensor Data

Jae Chan Jeong, Jun Soo Kim, Matthew P. Buman, Pavan Turaga, Eun Som Jeon

IEEE Internet of Things Journal · 2026-11 · 13 (22) · 55610-55628

DOI: 10.1109/jiot.2026.3733757 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Internet of Things Journal, 2026-11. Verfasst von Jae Chan Jeong und 4 weiteren Autor:innen. Die Arbeit nennt 64 Literaturangaben. Gefördert von: National Research Foundation of Korea (NRF) through Korean Government [Ministry of Science and ICT (MSIT)], NSF.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Wearable sensors are key components of the Internet of Things (IoT), enabling continuous and context-aware monitoring of human activities. As the demand for wearable sensors increases, accurate analysis of time-series data has become critical for reliable activity classification. However, raw wearable sensor signals are often affected by noise and have limited ability to capture contextual and discriminative information. To address these challenges, recent studies have investigated image representations derived from time-series data, such as persistence images (PIs) and Gramian angular fields (GAFs), which provide richer feature representations for learning. Meanwhile, wearable devices operate under stringent computational and memory constraints, making lightweight models essential for practical deployment. Knowledge distillation (KD) has therefore adopted as an effective approach to transfer knowledge from large, complex models to compact ones suitable for resource-constrained environments. Despite the representational advantages of image-based approaches, their effective integration into KD frameworks for wearable sensor data remains underexplored. In this article, we analyze how image-based representations contribute to activity classification from wearable sensor signals, and investigate how their incorporation into KD frameworks interacts with model compactness and representation richness. We further extend this analysis to the joint use of multiple image-based representations within KD, investigating their interactions and distillation behaviors. We conduct a comprehensive empirical study and extensive experimental analyses across multiple wearable sensor datasets, evaluating different KD strategies in terms of noise robustness, generalizability, intermediate feature representations, and computational efficiency. The results provide practical design guidelines to design compact and high-performance IoT-based wearable sensing systems.

Förderung

  • National Research Foundation of Korea (NRF) through Korean Government [Ministry of Science and ICT (MSIT)]
  • NSF

Zitieren

APA

Jeong, J. C., Kim, J. S., Buman, M. P., Turaga, P., & Jeon, E. S. (2026). Leveraging Multi-Image Representations for Improved Time-Series Knowledge Distillation on Wearable Sensor Data. IEEE Internet of Things Journal, 13(22), 55610-55628. https://doi.org/10.1109/jiot.2026.3733757

Vancouver

Jeong JC, Kim JS, Buman MP, Turaga P, Jeon ES. Leveraging Multi-Image Representations for Improved Time-Series Knowledge Distillation on Wearable Sensor Data. IEEE Internet of Things Journal. 2026;13(22):55610-55628. doi:10.1109/jiot.2026.3733757

BibTeX
@article{Jeong2026,
  author = {Jeong, Jae Chan and Kim, Jun Soo and Buman, Matthew P. and Turaga, Pavan and Jeon, Eun Som},
  title = {Leveraging Multi-Image Representations for Improved Time-Series Knowledge Distillation on Wearable Sensor Data},
  journal = {IEEE Internet of Things Journal},
  year = {2026},
  volume = {13},
  number = {22},
  pages = {55610--55628},
  doi = {10.1109/jiot.2026.3733757}
}
Kostenlos starten →