Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Video

Updated: Jun 13, 2026

Design and Analysis for Fall Detection System Simplification
08:05

Design and Analysis for Fall Detection System Simplification

Published on: April 6, 2020

Efficient Fall Detection from Wrist-Worn IMU Signals via Knowledge Distillation: A Lightweight CNN Approach Using the

Ali Taheri1, Mina Salehi2, Jeong Ho Kim3

  • 1Department of Electrical and Computer Engineering, Texas A&M University, College Station, TX 77843, USA.

Sensors (Basel, Switzerland)
|June 12, 2026
PubMed
Summary

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Perspectives of Mining Personnel on Adopting Occupational Exoskeletons: Comparisons Between a Developed and a Developing Country.

Mining, metallurgy & exploration·2025
See all related articles
This summary is machine-generated.

Knowledge distillation enhances wrist-worn inertial measurement unit (IMU) sensors for fall detection. This method improves accuracy and F1 scores for lightweight models, making fall detection more practical for wearable devices.

Area of Science:

  • Biomedical Engineering
  • Computer Science
  • Gerontology

Background:

  • Falls are a significant cause of injury and death in older adults.
  • Wearable inertial measurement unit (IMU) sensors offer potential for fall detection but often require multiple sensors and complex models.
  • Existing methods face limitations in practicality for resource-constrained wearable devices.

Purpose of the Study:

  • To propose and evaluate a knowledge distillation framework for efficient wrist-based fall detection.
  • To leverage a teacher-student model to improve the performance of a compact, wrist-only convolutional neural network (CNN).
  • To assess the feasibility of using knowledge distillation for fall detection on wearable devices.

Main Methods:

  • Implemented a knowledge distillation framework using a teacher-student model.
Keywords:
UMAFallfall detectionknowledge distillationlightweight CNNwearable sensorswrist-worn IMU

More Related Videos

Using Motion Capture Technology in the Instrumented Timed Up and Go Test to Detect the Risk of Falling in Aged Adults
05:26

Using Motion Capture Technology in the Instrumented Timed Up and Go Test to Detect the Risk of Falling in Aged Adults

Published on: October 25, 2024

Artificial Intelligence-Based System for Detecting Attention Levels in Students
06:37

Artificial Intelligence-Based System for Detecting Attention Levels in Students

Published on: December 15, 2023

Related Experiment Videos

Last Updated: Jun 13, 2026

Design and Analysis for Fall Detection System Simplification
08:05

Design and Analysis for Fall Detection System Simplification

Published on: April 6, 2020

Using Motion Capture Technology in the Instrumented Timed Up and Go Test to Detect the Risk of Falling in Aged Adults
05:26

Using Motion Capture Technology in the Instrumented Timed Up and Go Test to Detect the Risk of Falling in Aged Adults

Published on: October 25, 2024

Artificial Intelligence-Based System for Detecting Attention Levels in Students
06:37

Artificial Intelligence-Based System for Detecting Attention Levels in Students

Published on: December 15, 2023

  • The teacher model utilized IMU data from four body locations (waist, wrist, ankle, chest).
  • A compact, wrist-only CNN student model was trained using soft targets from the teacher model on the UMAFall dataset.

    • Main Results:

    • The teacher model achieved 97.6% accuracy and 96.7% F1 score.
    • The independently trained wrist-only CNN achieved 90.2% accuracy and 87.1% F1 score.
    • After knowledge distillation, the student model improved to 95.1% accuracy and 93.3% F1 score, maintaining a lightweight architecture.

    Conclusions:

    • Knowledge distillation significantly improves the performance of wrist-only fall detection models.
    • This approach offers a practical solution for fall detection in resource-constrained wearable devices.
    • Further validation with older adults and real-world smartwatch data is recommended.