EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Image Representation-Driven Knowledge Distillation for Improved Time-Series Interpretation on Wearable Sensor Data

Jae Chan Jeong iD, Matthew P Buman iD, Pavan Turaga iD, Eun Som Jeon iD

Sensors · 16.10.2025 · 25 (20) · 6396

DOI: 10.3390/s25206396 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Sensors, online erschienen am 16.10.2025. Verfasst von Jae Chan Jeong und 3 weiteren Autor:innen. Beteiligte Einrichtung: Department of Computer Science and Engineering, Seoul National University of Science and Technology, Seoul 01811, Republic of Korea. Hauptthemen laut MEDLINE: Wearable Electronic Devices, Image Processing, Computer-Assisted. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 58 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Seoul National University of Science and Technology.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

With the increased demand for wearable sensors, image representations—such as persistence images and Gramian angular fields—transformed from time-series data have been investigated to address challenges in wearables arising from physiological variations, sensor noise, and limitations in capturing contextual information. To preserve the lightweight structural design of models, knowledge distillation (KD) has also been employed alongside image representations during training to distill smaller and more efficient models. Although image representations play a key role in providing richer and more informative features in training a model, their effectiveness within the KD framework has not been thoroughly explored. In this paper, we focus on image representation-driven KD to investigate whether these representations can provide useful knowledge leading to improved time-series interpretation in activity classification tasks. We explore the benefits of integrating image representations into KD, and we analyze the interplay between representation richness and model compactness with different combinations of teacher and student networks. We also introduce diverse KD strategies to utilize image representations, and we demonstrate the strategies with various perspectives, such as analysis of noises, generalizability, and compatibility, across datasets of varying scales to obtain comprehensive and insightful observations. These offer valuable insights for designing efficient and high-performance wearable sensor-based systems.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Wearable Electronic DevicesImage Processing, Computer-AssistedHumansAlgorithms

Stichwörter der Autor:innen

Gramian angular fieldimage representationknowledge distillationpersistence imagetime-series data analysiswearable sensor data

Autor:innen und Institutionen

  • Department of Computer Science and Engineering, Seoul National University of Science and Technology, Seoul 01811, Republic of Korea.
  • College of Health Solutions, Arizona State University, Phoenix, AZ 85004, USA.
  • Geometric Media Lab, The GAME School, Arizona State University, Tempe, AZ 85281, USA.

ORCID-Profile: Jae Chan Jeong, Matthew P Buman, Pavan Turaga, Eun Som Jeon

Förderung

  • Seoul National University of Science and Technology

Zitieren

APA

Jeong, J. C., Buman, M. P., Turaga, P., & Jeon, E. S. (2025). Image Representation-Driven Knowledge Distillation for Improved Time-Series Interpretation on Wearable Sensor Data. Sensors, 25(20), 6396. https://doi.org/10.3390/s25206396

Vancouver

Jeong JC, Buman MP, Turaga P, Jeon ES. Image Representation-Driven Knowledge Distillation for Improved Time-Series Interpretation on Wearable Sensor Data. Sensors. 2025;25(20):6396. doi:10.3390/s25206396

BibTeX
@article{Jeong2025,
  author = {Jeong, Jae Chan and Buman, Matthew P and Turaga, Pavan and Jeon, Eun Som},
  title = {Image Representation-Driven Knowledge Distillation for Improved Time-Series Interpretation on Wearable Sensor Data},
  journal = {Sensors},
  year = {2025},
  volume = {25},
  number = {20},
  pages = {6396},
  doi = {10.3390/s25206396}
}
Kostenlos starten →