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 Concept Videos

You might also read

Related Articles

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

Sort by
Same author

Heat stress induced testicular impairment is related to orchitis and complement activation in Rongchang boars.

Journal of animal science and biotechnology·2025
Same author

Machine learning-assisted rapid screening of oil types on microfluidic thread-based analytical devices (μTADs).

Analytica chimica acta·2025
Same author

Nicotinamide Riboside Supplementation Alleviates Heat Stress-Induced Testicular Damage and Improves Semen Quality in Boars.

Molecular nutrition & food research·2025
Same author

Pairwise hemorrhage-brain region interaction-driven hemorrhagic stroke assessment in CT.

Physics in medicine and biology·2024
Same author

Exploring the impact of APOE ɛ4 on functional connectivity in Alzheimer's disease across cognitive impairment levels.

NeuroImage·2024
Same author

Multi-phase features interaction transformer network for liver tumor segmentation and microvascular invasion assessment in contrast-enhanced CT.

Mathematical biosciences and engineering : MBE·2024

Related Experiment Video

Updated: Jul 24, 2025

Sit-to-stand-and-walk from 120% Knee Height: A Novel Approach to Assess Dynamic Postural Control Independent of Lead-limb
08:24

Sit-to-stand-and-walk from 120% Knee Height: A Novel Approach to Assess Dynamic Postural Control Independent of Lead-limb

Published on: August 30, 2016

10.3K

Combining 3D skeleton data and deep convolutional neural network for balance assessment during walking.

Xiangyuan Ma1, Buhui Zeng1, Yanghui Xing1

  • 1Department of Biomedical Engineering, Shantou University, Shantou, China.

Frontiers in Bioengineering and Biotechnology
|July 6, 2023
PubMed
Summary

This study introduces a new automated method using 3D skeleton data and deep convolutional neural networks (DCNNs) for accurate balance assessment during walking. The DCNN approach significantly improves detection of balance impairment, aiding early disease intervention.

Keywords:
Kinectbalance assessmentdeep convolutional neural networkmachine learningskeleton data

More Related Videos

A Human-machine-interface Integrating Low-cost Sensors with a Neuromuscular Electrical Stimulation System for Post-stroke Balance Rehabilitation
11:06

A Human-machine-interface Integrating Low-cost Sensors with a Neuromuscular Electrical Stimulation System for Post-stroke Balance Rehabilitation

Published on: April 12, 2016

10.5K
Author Spotlight: Developing Low-Tech Balance Assessment Methods for Broad-Spectrum Healthcare Applications
06:28

Author Spotlight: Developing Low-Tech Balance Assessment Methods for Broad-Spectrum Healthcare Applications

Published on: September 1, 2023

4.8K

Related Experiment Videos

Last Updated: Jul 24, 2025

Sit-to-stand-and-walk from 120% Knee Height: A Novel Approach to Assess Dynamic Postural Control Independent of Lead-limb
08:24

Sit-to-stand-and-walk from 120% Knee Height: A Novel Approach to Assess Dynamic Postural Control Independent of Lead-limb

Published on: August 30, 2016

10.3K
A Human-machine-interface Integrating Low-cost Sensors with a Neuromuscular Electrical Stimulation System for Post-stroke Balance Rehabilitation
11:06

A Human-machine-interface Integrating Low-cost Sensors with a Neuromuscular Electrical Stimulation System for Post-stroke Balance Rehabilitation

Published on: April 12, 2016

10.5K
Author Spotlight: Developing Low-Tech Balance Assessment Methods for Broad-Spectrum Healthcare Applications
06:28

Author Spotlight: Developing Low-Tech Balance Assessment Methods for Broad-Spectrum Healthcare Applications

Published on: September 1, 2023

4.8K

Area of Science:

  • Biomechanics
  • Medical Imaging
  • Artificial Intelligence

Background:

  • Balance impairment is a key indicator for various diseases, necessitating early detection to mitigate fall risks and disease progression.
  • Current balance assessments rely on subjective evaluations using balance scales, limiting objectivity and precision.
  • Objective and automated methods are needed to improve the accuracy and timeliness of balance ability assessments.

Purpose of the Study:

  • To develop and validate an automated method for assessing walking balance abilities using 3D skeleton data and deep convolutional neural networks (DCNNs).
  • To compare the performance of the proposed DCNN method against traditional machine learning and other CNN-based approaches.
  • To identify the most relevant body parts for accurate balance assessment and interpret the DCNN model's decision-making process.

Main Methods:

  • A novel method combining 3D skeleton data and DCNNs was designed for automated balance assessment during walking.
  • A 3D skeleton dataset comprising three standardized balance ability levels was collected for model training and validation.
  • Extensive comparisons involving different skeleton-node selections, DCNN hyperparameters, and leave-one-subject-out cross-validation were conducted.

Main Results:

  • The proposed DCNN method achieved high accuracy (93.33%), precision (94.44%), and F1 score (94.46%), outperforming existing methods.
  • Analysis indicated that data from the body trunk and lower limbs are crucial for balance assessment, while upper limb data may decrease accuracy.
  • The DCNN model demonstrated improved accuracy in walking balance assessment compared to a state-of-the-art posture classification method.

Conclusions:

  • The developed DCNN-based approach offers a fast, accurate, and objective method for assessing walking balance abilities.
  • This automated system has the potential to significantly aid in the early detection and management of balance impairments.
  • Further research utilizing Layer-wise Relevance Propagation (LRP) can enhance the interpretability and clinical applicability of DCNN models in balance assessment.