Related Experiment Video
Updated: Oct 7, 2025
07:52
Automated Rat Single-Pellet Reaching with 3-Dimensional Reconstruction of Paw and Digit Trajectories
Published on: July 10, 2019
14.5K
Human Motion Tracking Using 3D Image Features with a Long Short-Term Memory Mechanism Model-An Example of Forward
Kai-Yu Chen1, Li-Wei Chou2, Hui-Min Lee3
1Department of Biomedical Engineering, National Yang Ming Chiao Tung University, Taipei 112, Taiwan.
Sensors (Basel, Switzerland)
|January 11, 2022
Summary
This study introduces a deep learning model for 3D human motion tracking using 3D image features. The new system offers improved accuracy over 2D systems for rehabilitation and other applications.
Area of Science:
- Biomedical Engineering
- Computer Vision
- Machine Learning
Background:
- Human motion tracking is crucial for rehabilitation, often using Inertial Measurement Units (IMUs) or 2D RGB imaging.
- IMUs offer accuracy but can be uncomfortable, while 2D systems lack spatial depth for precise 3D motion analysis.
Purpose of the Study:
- To develop an advanced human motion tracking technology utilizing 3D image features.
- To overcome the limitations of existing 2D visual-based and IMU-based tracking systems.
Main Methods:
- A deep learning (DL) model incorporating a deep bidirectional long short-term memory (DBLSTM) mechanism was employed.
- The system processes 3D image data to capture complex spatial changes in human movement.
Main Results:
- The proposed 3D DL system demonstrated superior human motion tracking performance compared to traditional 2D systems.
- Achieved a Root Mean Square Error (RMSE) below 0.5 m/s² in acceleration across X, Y, and Z directions.
Conclusions:
- The DBLSTM-based 3D human motion tracking model presents a viable and accurate alternative for various applications.
- This technology holds promise for enhancing future human motion analysis in fields like rehabilitation.
