A comparison of currently available optoelectronic motion capture systems
Introduction
Whole body human motion is commonly quantified utilizing commercially available, passive, retroreflective marker based, optoelectronic motion capture systems. The measurement of human motion relies heavily on the ability of the optoelectronic motion capture systems to accurately measure retroreflective marker positions. Marker position accuracy is largely dependent on the optical characteristics of the camera system and algorithms implemented in the tracking software. In the past, researchers have intermittently reported on the characteristics of individual motion capture systems and in some cases, their relative accuracy (Van Der Kruk and Reijne, 2018, Windolf et al., 2008). In 1999, Richards expanded camera evaluation by critically reviewing multiple camera systems and their ability to generate the same marker coordinate positions under different field tests (Richards, 1999). The project results indicated that some systems outperformed others.
Field tests were designed utilizing what became to be known as the Standard Assessment of Motion System Accuracy (SAMSA) device shown in Fig. 1 (Richards, 1999). The SAMSA device has seven fixed markers at known distances from each other. For each field test, the camera systems generated coordinate data for each marker during a four second capture at 60 Hz. Distances between and angles formed by the recorded coordinate positions were then compared to known distances and angles between marker groups. The field tests provided the accuracy and precision of the era’s prominent camera systems. Average absolute measurement errors associated with marker distances exceeded one millimeter for distance tests with RMSE values exceeding one millimeter for every camera system. In addition, some systems recorded a centimeter of error for at least one frame for several of the marker distance measurements tested.
In the subsequent 20 years, motion capture technology has significantly improved. Camera resolutions are now typically over one megapixel (MP) and manufacturer tracking software allows for marker replacement/switching, gap filling, and signal filtering. The purpose of this study was to replicate the tests from 1999 using current, higher-resolution motion capture systems and current calibration techniques. A clinical approach was taken during this investigation to reflect the variation in lab organizations across institutions. Specifically, systems were compared in their permanent working environments to determine if results were analogous across laboratories. The review produced characteristics and specifications of each system and their measurement performances in a similar manner as the 1999 Richards investigation (Richards, 1999).
Section snippets
Manufacturer specifications
Table 1 provides the manufacturer specifications for each system tested. All manufacturer software was able to gap fill and switch mislabeled markers.
SAMSA device description
Marker coordinate data was obtained utilizing the original prototype SAMSA device from the 1999 study (Fig. 1). Seven markers are permanently fixed onto the device. Markers one and two are affixed to the superior surface of a rigid aluminum arm that rotates 360˚ in the transverse plane of the capture volume at 60 rpm. Both marker one and two were visible to all cameras throughout the 360˚ rotation. An aluminum plate was attached vertically to the end of the rotating arm so that the face of the
Results
The distance measurement errors for the top two rotating arm markers are shown in Table 3. Every system was able to measure the average distance between the two markers with less than one-millimeter error with the greatest average difference measuring approximately 0.5 mm. The highest reported standard deviation was approximately 0.1 mm. Additionally, the maximum difference from the calculated FaroArm distance measured approximately 1.0 mm.
The distance measurement errors for the top two plate
Discussion
Along with higher camera resolutions, tracking software has improved since the 1999 investigation. The original investigation sought to assess the accuracy of available systems and the tracking software’s ability to recreate missing markers. All current systems were able to successfully track all seven markers without discontinuities throughout the trials including markers 6 & 7 before, during, and after contact. Occasionally, two markers would switch marker labels in one or two frames during a
Acknowledgements
We would like to thank all of the gait and motion analysis labs that allowed us to conduct this research, including: the BIOMS labs at the University of Delaware, The Human Performance Lab at Thomas Jefferson University, and the Motion Analysis Center at Shriners Hospital for Children – Philadelphia. No funding was provided for this study.
Conflict of Interest
The authors declare that they have no conflict of interest.
References (6)
- et al.
Optimal camera placement for accurate reconstruction
Pattern Recognition
(2002) The measurement of human motion: A comparison of commercially available systems
Human Movement Science
(1999)- et al.
Systematic accuracy and precision analysis of video motion capturing systems-exemplified on the Vicon-460 system
Journal of Biomechanics
(2008)
Cited by (88)
A comparison of three-dimensional kinematics between markerless and marker-based motion capture in overground gait
2023, Journal of BiomechanicsEffects of soft tissue artifacts on the calculated kinematics of the knee during walking and running
2023, Journal of BiomechanicsA review of combined functional neuroimaging and motion capture for motor rehabilitation
2024, Journal of NeuroEngineering and Rehabilitation