Comparative Study On Gait Kinematics Between Microsoft Kinect and Vicon Across Different Anthropometric Measurements

Sagida M.A. Bilal, Aizreena Binti Azaman

Abstract

The main challenge in gait analysis is to provide an accessible and affordable mean of recording a gait test. Currently, Kinect is presented as a new innovative approach in measuring gait parameters. But then, the use of Kinect camera is limited, noticing it is focused exclusively in particular groups especially adults. Therefore, a study that encompasses walking analysis in multiple anthropometry based populations can provide a comprehensive understanding of the ability of Kinect modality in capturing human motion. In this study, Kinect capability was examined in a number of ten subjects (5 adults and 5 children). Each subject was asked to walk an overall 4m distance. Vicon capturing system and Kinect camera were used to record the walking simultaneously. Kinematics data such as hip flexion angle and knee flexion angle were recorded. The results were compared using Pearson’s correlation coefficient and standard deviation. The results showed that a lower correlation between the two systems was found in terms of knee and hip flexion angles with the correlation ranging from r = -0.30 to r =0.932 in the right leg. similarly, the left leg correlation ranged from r = - 0.143 to r =0.70. Likewise, in order to examine the optimum lab setting, in terms of distance and angle position of the camera. The subjects were asked to walk to 1.5 m and then 3m distance with Kinect positioned at 90° and 75°. Unexpectedly, with the correlation results indicating that placing Kinect at 3m distance and angled at 75° harbored a better correlation. These findings suggest that in general Kinect is applicable in measuring gait Kinematics with the needed modification to the analyzing software and adjustment to the camera setting.