Methods for Spatiotemporal Analysis of Human Gait Based on Data from Depth Sensors

Abstract

Gait analysis may serve various purposes related to health care, such as the estimation of elderly people's risk of falling. This paper is devoted to gait analysis based on data from depth sensors which are suitable for use both at healthcare facilities and in monitoring systems dedicated to household environments. This paper is focused on the comparison of three methods for spatiotemporal gait analysis based on data from depth sensors, involving the analysis of the movement trajectories of the knees, feet, and centre of mass. The accuracy of the results obtained using those methods was assessed for different depth sensors' viewing angles and different types of subject clothing. Data were collected using a Kinect v2 device. Five people took part in the experiments. Data from a Zebris FDM platform were used as a reference. The obtained results indicate that the viewing angle and the subject's clothing affect the uncertainty of the estimates of spatiotemporal gait parameters, and that the method based on the trajectories of the feet yields the most information, while the method based on the trajectory of the centre of mass is the most robust.

Keywords: data processing; depth sensor; gait analysis; health care; in-home monitoring.

Figure 1

(a) An exemplary depth image acquired using a Kinect v2 device, in which brighter pixels indicate larger distances from the device; (b) The results of the detection of a human silhouette and the results of the localisation of 21 anatomical landmarks, obtained using the algorithm implemented in the Kinect v2 device.

Figure 2

(a) The coordinate system (x_p, y_p, z_p) associated with the examined person, in which the x_p axis corresponds to that person’s medio-lateral axis, the y_p axis—to that person’s vertical axis and the z_p axis—to that person’s anteroposterior axis; (b) The angle ϑ of rotation necessary to make one of the coordinate axes associated with the depth sensor parallel to the examined person’s vertical axis; (c) The angle φ of rotation necessary to make one of the coordinate axes associated with the depth sensor parallel to the examined person’s walking direction.

Figure 3

(a) Exemplary estimates of anteroposterior positions of knees during walking, obtained using a depth sensor, and the results of their smoothing using a Savitzky-Golay filter; (b) An exemplary dependence of the anteroposterior distance between a walking person’s knees on time, with its local extrema indicated with circles and the silhouettes of that person extracted from the corresponding depth images.

Figure 4

(a) Exemplary estimates of the height of the examined person’s centre of mass during walking, obtained using a depth sensor, and the results of their smoothing using a Savitzky-Golay filter; (b) an exemplary dependence of the height of a walking person’s centre of mass on time, with its local extrema indicated with circles and the corresponding silhouettes of that person extracted from the depth images.

Figure 5

(a) Exemplary estimates of anteroposterior positions of a walking person’s feet, obtained using a depth sensor, and the results of their smoothing using a Savitzky-Golay filter; (b) Exemplary dependences of the horizontal velocity of a walking person’s feet on time, with the empirically selected threshold value indicated by the horizontal dashed line, the detected FO and FC moments indicated by the vertical dashed lines, and the silhouettes of that person extracted from the corresponding depth images.

Figure 6

Examples of the results of detecting the silhouettes and locating the anatomical landmarks of the 5 people participating the experiment.

Figure 7

Configurations of the devices used in the experiments in which the angle φ between the depth sensor’s line of sight and the walking trajectory is approximately (a) 180°, (b) 135° and (c) 90°.

Figure 8

The distribution of the errors of the estimates of the step time (a,b) and step length (c,d), obtained using the studied data processing methods; the figures on the left (a,c) present the results obtained for subjects wearing typical trousers, walking towards the depth sensor; the figures on the right (b,d) present the results obtained for all subjects in all experiments; the horizontal lines indicate the ME values; the height of each box equals the SDE value; the whiskers indicate the maximum and minimum errors; the black boxes represent the dispersion of the reference values obtained using the Zebris FDM platform.

Figure 9

The values of MARE obtained for the step time (a) and step length (b) for Subjects #1–#3 for different angles φ between the walking direction and the depth sensor’s line of sight.

Figure 10

The values of MARE obtained for the step time (a) and step length (b) for φ = 90°, 135°, 180° for each subject; the numbers above each bar indicate the subject.