Related Experiment Videos
The three-dimensional measurement of unconstrained motion using a model-matching method
1Faculty of Science and Technology, Keio University, Yokohama-shi, Japan.
Ergonomics
|March 1, 1994
Summary
This study introduces a novel image processing technique for accurate human motion measurement. It overcomes limitations of traditional methods, enabling unconstrained 3D motion analysis with high precision.
Area of Science:
- Biomechanics
- Computer Vision
- Human Motion Analysis
Background:
- Traditional motion measurement methods like goniometry and surface markers have limitations, including motion constraint and marker loss.
- Quantitative mechanical analysis of human movement relies on accurate motion measurement techniques.
Purpose of the Study:
- To develop a new, unconstrained motion measurement technique using image processing.
- To reduce processing time and improve accuracy in human motion analysis.
Main Methods:
- A geometric model of the human body is fitted to image sequences of whole-body movements.
- Processing time was reduced by utilizing overlap area information between models and images, instead of contour information.
- Joint angles of hidden segments were estimated using grayscale image information.
Main Results:
- The new method eliminates the need for segment coloring and surface markers.
- Three-dimensional finger motion (21 segments) was measured on a personal computer in 3-5 minutes per frame.
- The maximum error in joint angle estimation was 2 degrees.
Conclusions:
- The proposed image processing technique enables unconstrained human motion measurement with high accuracy and efficiency.
- This method significantly reduces processing time compared to previous image-based approaches.
- The technique is suitable for detailed 3D motion analysis, including complex movements like finger articulation.