Joint Position Registration between OpenPose and Motion Analysis for Rehabilitation

We developed a registration system between OpenPose and Motion Analysis, that is a gold standard of human motion analysis, to evaluate the accuracy of human joint position. We took the data of the both for walking ten healthy persons: five males and five females. The registration result was evaluate...

Full description

Saved in:
Bibliographic Details
Published in2020 IEEE 50th International Symposium on Multiple-Valued Logic (ISMVL) pp. 100 - 104
Main Authors Otsuka, Kazuki, Yagi, Naomi, Yamanaka, Yuki, Hata, Yutaka, Sakai, Yoshitada
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.11.2020
Subjects
Cameras
Correlation coefficient
Hip
human motion analysis
Knee
Legged locomotion
Motion Analysis
OpenPose
registration
rehabilitation
Shoulder
Wrist
Online AccessGet full text

Cover

More Information
Summary:We developed a registration system between OpenPose and Motion Analysis, that is a gold standard of human motion analysis, to evaluate the accuracy of human joint position. We took the data of the both for walking ten healthy persons: five males and five females. The registration result was evaluated by comparing with their obtained data. As a result, OpenPose had an error of 63.4mm on horizontal axis, 30.2mm on vertical axis in videos taken from the side view, and 29.2mm on horizontal axis, 34.1mm on vertical axis in videos taken from the front view with comparison to Motion Analysis.
ISSN:2378-2226
DOI:10.1109/ISMVL49045.2020.00-22