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

Protocol for longitudinal two-photon calcium imaging and holographic optogenetic manipulation to investigate memory in mice.

STAR protocols·2026
Same author

Beclin-1/LC3B regulation of podocyte autophagy in IgA nephropathy via PI3K/AKT signaling.

Journal of nephrology·2026
Same author

Deep learning-based real-time intraoperative detection of thoracic duct.

Journal of thoracic disease·2026
Same author

Association Between Extreme Climate Exposures and Bodily Pain: Evidence from the Chinese Health and Retirement Longitudinal Study.

Journal of epidemiology and global health·2026
Same author

Rapid detection of porcine circovirus type2 using a combination of RAA isothermal amplification and Cas12a nucleic detection technology.

Research in veterinary science·2026
Same author

Exploration of the viewing intention mechanism of Chinese traditional culture animated films based on SOR and TPB.

Frontiers in psychology·2026

Related Experiment Video

Updated: Nov 19, 2025

Capturing Dynamic Finger Gesturing with High-resolution Surface Electromyography and Computer Vision
08:15

Capturing Dynamic Finger Gesturing with High-resolution Surface Electromyography and Computer Vision

Published on: March 28, 2025

966

Wearable on-device deep learning system for hand gesture recognition based on FPGA accelerator.

Weibin Jiang1, Xuelin Ye2, Ruiqi Chen1,3

  • 1College of Physics and Information Engineering, Fuzhou University, Fuzhou 350116, China.

Mathematical Biosciences and Engineering : MBE
|February 2, 2021
PubMed
Summary

This study introduces a wearable deep learning system for on-device gesture recognition using an FPGA accelerator. The system achieves high accuracy (97%) with low power and latency, outperforming traditional cloud-based methods for applications like rehabilitation.

Keywords:
acceleratorconvolutional neural network (CNN)field-programmable gate array (FPGA)gesture recognitioninertial measurement unit (IMU)micro-control unit (MCU)

More Related Videos

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention
06:37

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention

Published on: December 15, 2023

4.8K
Author Spotlight: Enhancing Grasping Abilities for Hemiplegic Patients with Flexible Robotic Limbs
03:55

Author Spotlight: Enhancing Grasping Abilities for Hemiplegic Patients with Flexible Robotic Limbs

Published on: October 27, 2023

2.5K

Related Experiment Videos

Last Updated: Nov 19, 2025

Capturing Dynamic Finger Gesturing with High-resolution Surface Electromyography and Computer Vision
08:15

Capturing Dynamic Finger Gesturing with High-resolution Surface Electromyography and Computer Vision

Published on: March 28, 2025

966
Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention
06:37

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention

Published on: December 15, 2023

4.8K
Author Spotlight: Enhancing Grasping Abilities for Hemiplegic Patients with Flexible Robotic Limbs
03:55

Author Spotlight: Enhancing Grasping Abilities for Hemiplegic Patients with Flexible Robotic Limbs

Published on: October 27, 2023

2.5K

Area of Science:

  • Computer Science
  • Electrical Engineering
  • Biomedical Engineering

Background:

  • Gesture recognition is vital for Human-Computer Interaction (HCI) in fields like healthcare and rehabilitation.
  • Current methods often rely on cloud processing, posing challenges for real-time applications and remote monitoring.
  • Inertial Measurement Unit (IMU) sensors are commonly used for data collection.

Purpose of the Study:

  • To develop a low-power, low-latency wearable deep learning system for on-device gesture recognition.
  • To enable local data processing, enhancing convenience for applications such as remote rehabilitation training.
  • To present a novel software-hardware co-design approach for edge devices.

Main Methods:

  • Utilized a Field-Programmable Gate Array (FPGA) accelerator and a Cortex-M0 IP core for local processing.
  • Implemented a pre-stage processing module and a serial-parallel hybrid method for efficiency.
  • Employed a Convolutional Neural Network (CNN) and a Multilayer Perceptron (NN) for feature extraction and gesture classification.

Main Results:

  • The wearable system achieves high gesture recognition accuracy of 97%.
  • Demonstrated low-power and low-latency performance at the Micro Control Unit (MCU) level.
  • Outperformed Single Board Computers (SBCs), with performance more than double that of Cortex-A53 processors.

Conclusions:

  • The proposed wearable deep learning system offers an efficient solution for on-device gesture recognition.
  • The software-hardware co-design method provides a valuable reference for developing edge devices in various scenarios.
  • This approach enhances the feasibility of real-time gesture analysis in healthcare and other HCI applications.