Anatomical calibration for wearable motion capture systems: Video calibrated anatomical system technique

作者：Bisi Maria Cristina^*; Stagni Rita; Caroselli Alessio; Cappello Angelo

来源：Medical Engineering & Physics, 2015, 37(8): 813-819.

DOI：10.1016/j.medengphy.2015.05.013

• Human movement analysis
• Joint kinematics
• Anatomical calibration
• Wearable sensors

Inertial sensors are becoming widely used for the assessment of human movement in both clinical and research applications, thanks to their usability out of the laboratory. This work aims to propose a method for calibrating anatomical landmark position in the wearable sensor reference frame with an ease to use, portable and low cost device.
An off-the-shelf camera, a stick and a pattern, attached to the inertial sensor, compose the device. The proposed technique is referred to as video Calibrated Anatomical System Technique (vCAST). The absolute orientation of a synthetic femur was tracked both using the vCAST together with an inertial sensor and using stereo-photogrammetry as reference.
Anatomical landmark calibration showed mean absolute error of 0.6 +/- 0.5 mm: these errors are smaller than those affecting the in-vivo identification of anatomical landmarks. The roll, pitch and yaw anatomical frame orientations showed root mean square errors close to the accuracy limit of the wearable sensor used (1 degrees), highlighting the reliability of the proposed technique.
In conclusion, the present paper proposes and preliminarily verifies the performance of a method (vCAST) for calibrating anatomical landmark position in the wearable sensor reference frame: the technique is low time consuming, highly portable, easy to implement and usable outside laboratory.

• 出版日期2015-8