Concurrent validity of human pose tracking in video for measuring gait parameters in older adults: a preliminary analysis with multiple trackers, viewing angles, and walking directions

Sina Mehdizadeh, Hoda Nabavi, Andrea Sabo, Twinkle Arora, Andrea Iaboni, Babak Taati, Sina Mehdizadeh, Hoda Nabavi, Andrea Sabo, Twinkle Arora, Andrea Iaboni, Babak Taati

Abstract

Background: Many of the available gait monitoring technologies are expensive, require specialized expertise, are time consuming to use, and are not widely available for clinical use. The advent of video-based pose tracking provides an opportunity for inexpensive automated analysis of human walking in older adults using video cameras. However, there is a need to validate gait parameters calculated by these algorithms against gold standard methods for measuring human gait data in this population.

Methods: We compared quantitative gait variables of 11 older adults (mean age = 85.2) calculated from video recordings using three pose trackers (AlphaPose, OpenPose, Detectron) to those calculated from a 3D motion capture system. We performed comparisons for videos captured by two cameras at two different viewing angles, and viewed from the front or back. We also analyzed the data when including gait variables of individual steps of each participant or each participant's averaged gait variables.

Results: Our findings revealed that, i) temporal (cadence and step time), but not spatial and variability gait measures (step width, estimated margin of stability, coefficient of variation of step time and width), calculated from the video pose tracking algorithms correlate significantly to that of motion capture system, and ii) there are minimal differences between the two camera heights, and walks viewed from the front or back in terms of correlation of gait variables, and iii) gait variables extracted from AlphaPose and Detectron had the highest agreement while OpenPose had the lowest agreement.

Conclusions: There are important opportunities to evaluate models capable of 3D pose estimation in video data, improve the training of pose-tracking algorithms for older adult and clinical populations, and develop video-based 3D pose trackers specifically optimized for quantitative gait measurement.

Keywords: Deep learning; Gait; Human pose estimation; Walking.

Conflict of interest statement

The authors have no conflicts to report.

© 2021. The Author(s).

Figures

Fig. 1
Fig. 1
The pose tracking skeletons overlaid on the eye-level camera video for front (top row) and back (bottom row) view walks, A AlphaPose, B OpenPose, C Detectron
Fig. 2
Fig. 2
The pose tracking skeletons overlaid on the top camera video for front (top row) and back (bottom row) view walks, A AlphaPose, B OpenPose, C Detectron
Fig. 3
Fig. 3
Scatter plots for one temporal (step time, left column), and one spatial (step width, right column) gait variable, including each individual step captured in the front view walking bout with the eye-level camera, A AlphaPose, B OpenPose, C Detectron. The colors are associated with different participants’ data. The thick red line is the fitted line. The correlation values are also shown in the figures
Fig. 4
Fig. 4
the Bland–Altman plots between pairs of the pose tracking algorithms for the step width from the videos of the top (A) and eye-level (B) cameras. The top row in each panel is for the front view walks and the bottom row is for the back view walks. The dashed lines are the lower and upper limits of agreement (1.96*standard deviation) as well as the zero line
Fig. 5
Fig. 5
the Bland–Altman plots between pairs of the pose tracking algorithms for the estimated margin of stability (eMOS) from the videos of the top (A) and eye-level (B) cameras. The top row in each panel is for the front view walks and the bottom row is for the back view walks. The dashed lines are the lower and upper limits of agreement (1.96*standard deviation) as well as the zero line

References

    1. Brodie MA, Coppens MJ, Ejupi A, Gschwind YJ, Annegarn J, Schoene D, et al. Comparison between clinical gait and daily-life gait assessments of fall risk in older people. Geriatr Gerontol Int. 2017;17(11):2274–2282. doi: 10.1111/ggi.12979.
    1. Latorre J, Colomer C, Alcañiz M, Llorens R. Gait analysis with the Kinect v2: normative study with healthy individuals and comprehensive study of its sensitivity, validity, and reliability in individuals with stroke. J Neuroeng Rehabil. 2019;16(1):97. doi: 10.1186/s12984-019-0568-y.
    1. Mehdizadeh S, Dolatabadi E, Ng K-D, Mansfield A, Flint A, Taati B, et al. Vision-based assessment of gait features associated with falls in people with dementia. J Gerontol. 2019;75(6):1148–1153. doi: 10.1093/gerona/glz187.
    1. Dolatabadi E, Taati B, Mihailidis A. An automated classification of pathological gait using unobtrusive sensing technology. IEEE Trans Neural Syst Rehabil Eng. 2017;25(12):2336–2346. doi: 10.1109/TNSRE.2017.2736939.
    1. Dolatabadi E, Taati B, Mihailidis A. Vision-based approach for long-term mobility monitoring: single case study following total hip replacement. J Rehabil Res Dev. 2014;51(7):1165–1176. doi: 10.1682/JRRD.2013.12.0263.
    1. Mehdizadeh S, Sabo A, Ng K-D, Mansfield A, Flint AJ, Taati B, et al. Predicting Short-Term Risk of Falls in a High-Risk Group With Dementia. Journal of the American Medical Directors Association. 2020.
    1. Springer S, Yogev SG. Validity of the kinect for gait assessment: a focused review. Sensors (Basel, Switzerland) 2016;16(2):194. doi: 10.3390/s16020194.
    1. Cao Z, Hidalgo G, Simon T, Wei S-E, Sheikh Y. OpenPose: realtime multi-person 2D pose estimation using part affinity fields. IEEE Trans Pattern Anal Mach Intell. 2021;43(1):172–186. doi: 10.1109/TPAMI.2019.2929257.
    1. Zago M, Luzzago M, Marangoni T, De Cecco M, Tarabini M, Galli M. 3D tracking of human motion using visual skeletonization and stereoscopic vision. Front Bioeng Biotechnol. 2020;8:181. doi: 10.3389/fbioe.2020.00181.
    1. Viswakumar A, Rajagopalan V, Ray T, Parimi C, editors. Human Gait Analysis Using OpenPose. 2019 Fifth International Conference on Image Information Processing (ICIIP); 2019 15–17 Nov. 2019.
    1. Kidziński Ł, Yang B, Hicks JL, Rajagopal A, Delp SL, Schwartz MH. Deep neural networks enable quantitative movement analysis using single-camera videos. Nat Commun. 2020;11(1):4054. doi: 10.1038/s41467-020-17807-z.
    1. Sato K, Nagashima Y, Mano T, Iwata A, Toda T. Quantifying normal and parkinsonian gait features from home movies: Practical application of a deep learning–based 2D pose estimator. PLOS ONE. 2019;14(11):e0223549. doi: 10.1371/journal.pone.0223549.
    1. Fang H-S, Xie S, Tai Y-W, Lu C. RMPE: Regional Multi-person Pose Estimation. arXiv:161200137 [cs]. 2018.
    1. Girshick R, Radosavovic I, Gkioxari G, Dollár P, He K. Detectron 2018 [Available from: .
    1. Stenum J, Rossi C, Roemmich RT. Two-dimensional video-based analysis of human gait using pose estimation. PLOS Comput Biol. 2021;17(4):e1008935. doi: 10.1371/journal.pcbi.1008935.
    1. Ng K-D, Mehdizadeh S, Iaboni A, Mansfield A, Flint A, Taati B. Measuring gait variables using computer vision to assess mobility and fall risk in older adults with dementia. IEEE J Transl Eng Health Med. 2020;8:1–9. doi: 10.1109/JTEHM.2020.2998326.
    1. Ota M, Tateuchi H, Hashiguchi T, Ichihashi N. Verification of validity of gait analysis systems during treadmill walking and running using human pose tracking algorithm. Gait Posture. 2021;85:290–297. doi: 10.1016/j.gaitpost.2021.02.006.
    1. Osoba MY, Rao AK, Agrawal SK, Lalwani AK. Balance and gait in the elderly: a contemporary review. Laryngoscope Invest Otolaryngol. 2019;4(1):143–153. doi: 10.1002/lio2.252.
    1. Folstein MF, Folstein SE, McHugh PR. "Mini-mental state". A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res. 1975;12(3):189–98. doi: 10.1016/0022-3956(75)90026-6.
    1. Sterke CS, Huisman SL, van Beeck EF, Looman CW, van der Cammen TJ. Is the Tinetti Performance Oriented Mobility Assessment (POMA) a feasible and valid predictor of short-term fall risk in nursing home residents with dementia? Int Psychogeriatr. 2010;22(2):254–263. doi: 10.1017/S1041610209991347.
    1. Berg K, Wood-Dauphine S, Williams JI, Gayton D. Measuring balance in the elderly: preliminary development of an instrument. Physiother Can. 1989;41(6):304–311. doi: 10.3138/ptc.41.6.304.
    1. Podsiadlo D, Richardson S. The timed "Up & Go": A test of basic functional mobility for frail elderly persons. J Am Geriatr Soc. 1991;39(2):142–148. doi: 10.1111/j.1532-5415.1991.tb01616.x.
    1. Al-Amri M, Nicholas K, Button K, Sparkes V, Sheeran L, Davies JL. Inertial measurement units for clinical movement analysis: reliability and concurrent validity. Sensors. 2018;18(3):719. doi: 10.3390/s18030719.
    1. Karatsidis A, Jung M, Schepers HM, Bellusci G, de Zee M, Veltink PH, et al. Musculoskeletal model-based inverse dynamic analysis under ambulatory conditions using inertial motion capture. Med Eng Phys. 2019;65:68–77. doi: 10.1016/j.medengphy.2018.12.021.
    1. Teufl W, Lorenz M, Miezal M, Taetz B, Fröhlich M, Bleser G. Towards inertial sensor based mobile gait analysis: event-detection and spatio-temporal parameters. Sensors (Basel, Switzerland) 2018;19(1):38. doi: 10.3390/s19010038.
    1. Pham TT, Moore ST, Lewis SJG, Nguyen DN, Dutkiewicz E, Fuglevand AJ, et al. Freezing of gait detection in parkinson's disease: a subject-independent detector using anomaly scores. IEEE Trans Biomed Eng. 2017;64(11):2719–2728. doi: 10.1109/TBME.2017.2665438.
    1. Hof AL. The ‘extrapolated center of mass’ concept suggests a simple control of balance in walking. Hum Mov Sci. 2008;27(1):112–125. doi: 10.1016/j.humov.2007.08.003.
    1. Hamacher D, Singh NB, Van Dieën JH, Heller MO, Taylor WR. Kinematic measures for assessing gait stability in elderly individuals: a systematic review. J R Soc Interface. 2011;8(65):1682–1698. doi: 10.1098/rsif.2011.0416.
    1. Ohtsu H, Yoshida S, Minamisawa T, Katagiri N, Yamaguchi T, Takahashi T, et al. Does the balance strategy during walking in elderly persons show an association with fall risk assessment? J Biomech. 2020;103:109657. doi: 10.1016/j.jbiomech.2020.109657.
    1. Sabo A, Mehdizadeh S, Ng K-D, Iaboni A, Taati B. Assessment of Parkinsonian gait in older adults with dementia via human pose tracking in video data. J Neuroeng Rehabil. 2020;17(1):97. doi: 10.1186/s12984-020-00728-9.
    1. Mukaka MM. Statistics corner: a guide to appropriate use of correlation coefficient in medical research. Malawi Med J. 2012;24(3):69–71.
    1. Lin T-Y, Maire M, Belongie S, Bourdev L, Girshick R, Hays J, et al. Microsoft COCO: Common Objects in Context. arXiv:14050312 [cs]. 2015.
    1. Seethapathi N, Wang S, Saluja R, Blohm G, Kording KP. Movement science needs different pose tracking algorithms. arXiv:190710226 [cs, q-bio]. 2019.
    1. Robinson WS. Ecological correlations and the behavior of individuals. Am Sociol Rev. 1950;15(3):351–357. doi: 10.2307/2087176.
    1. Guo Y, Deligianni F, Gu X, Yang G. 3-D canonical pose estimation and abnormal gait recognition with a single RGB-D camera. IEEE Robot Autom Lett. 2019;4(4):3617–3624. doi: 10.1109/LRA.2019.2928775.
    1. Chen C-H, Ramanan D. 3D Human Pose Estimation = 2D Pose Estimation + Matching. arXiv:161206524 [cs]. 2017.

Source: PubMed

Sponzoři a spolupracovníci

Zdravotní podmínky

Drogové intervence

3
Předplatit