Marker Set Configuration and Rigid Body Attitude Determination

Research output: Contribution to journalArticlepeer-review

Abstract

Image-based motion-analysis systems typically place markers on the bodies of interest. The error in determining segment attitude from these markers is a function the marker position errors, the number of markers, and the spatial distribution of the markers. The spatial distribution includes two factors: the mean square distance of these markers to their geometric center, and the degree of anisotropy in the marker distribution. The purposes of this study were to: (1) present a metric which quantifies the marker spatial distribution (anisotropic to isotropic) and (2) examine the influence of marker distribution on the accuracy of rigid body attitude determination. To test the influence of the marker distribution on body attitude determination 1000 criterion attitudes were determined. These attitudes then had to be estimated for two marker sets for which the marker distribution metric, noise levels, and root-mean-square distance of the markers were systematically varied. Anisotropic marker distributions were shown to negatively affect the accuracy of attitude determination. The influence of anisotropic marker distributions on attitude accuracy could be blunted by increasing the number of markers, increasing the root-mean-square distance of markers from their geometric center, and reducing noise levels. These results have implications for the measurement of the attitudes of body segments. For example, the ability to have a large spatial distribution of markers and a large number of markers to maximize the measurement accuracy of segment attitude is different for a small segment such as the fifth metacarpal compared with the thigh.

Original languageEnglish (US)
Article number124501
JournalJournal of Biomechanical Engineering
Volume145
Issue number12
DOIs
StatePublished - Dec 1 2023

All Science Journal Classification (ASJC) codes

  • Biomedical Engineering
  • Physiology (medical)

Cite this