Elsevier

Gait & Posture

Volume 51, January 2017, Pages 77-83
Gait & Posture

Full length article
Improved kinect-based spatiotemporal and kinematic treadmill gait assessment

https://doi.org/10.1016/j.gaitpost.2016.10.001Get rights and content

Highlights

  • Kinect v2 provides comparable data to a standard 3D motion analysis system.

  • Kinect v2 can be a clinical tool for evaluating hip and knee gait kinematics.

  • Kinect can be a clinical tool for evaluating sagittal plane spatiotemporal variables.

Abstract

A cost-effective, clinician friendly gait assessment tool that can automatically track patients’ anatomical landmarks can provide practitioners with important information that is useful in prescribing rehabilitative and preventive therapies. This study investigated the validity and reliability of the Microsoft Kinect v2 as a potential inexpensive gait analysis tool. Ten healthy subjects walked on a treadmill at 1.3 and 1.6 m·s−1, as spatiotemporal parameters and kinematics were extracted concurrently using the Kinect and three-dimensional motion analysis. Spatiotemporal measures included step length and width, step and stride times, vertical and mediolateral pelvis motion, and foot swing velocity. Kinematic outcomes included hip, knee, and ankle joint angles in the sagittal plane. The absolute agreement and relative consistency between the two systems were assessed using interclass correlations coefficients (ICC2,1), while reproducibility between systems was established using Lin’s Concordance Correlation Coefficient (rc). Comparison of ensemble curves and associated 90% confidence intervals (CI90) of the hip, knee, and ankle joint angles were performed to investigate if the Kinect sensor could consistently and accurately assess lower extremity joint motion throughout the gait cycle. Results showed that the Kinect v2 sensor has the potential to be an effective clinical assessment tool for sagittal plane knee and hip joint kinematics, as well as some spatiotemporal temporal variables including pelvis displacement and step characteristics during the gait cycle.

Introduction

Gait analysis is an effective clinical tool used for a wide range of applications including evaluating neurological diseases [1], [2], fall risk [3], orthopedic disability [4], and progress during rehabilitation [5]. While optoelectronic motion capture systems are the gold standard for dynamic movement assessment, the high financial cost and technical expertise required to operate these systems and analyze the data, make them an unrealistic option for clinical use. Although movement patterns are commonly used to assess injury risk, progress during rehabilitation, and functional performance, these assessments are often subjective limiting their effectiveness, reliability and sensitivity to change [6], [7]. The objective, quantitative nature of motion analysis could provide a clear improvement over these techniques and be of significant value within the clinical environment. Therefore, a publically available, cost-effective and clinician-friendly motion capture solution, allowing valid and reliable assessments of kinematic and spatiotemporal variables during functional movements, represents a logical step toward improved patient care.

Newly developed motion analysis technologies afford researchers and clinicians multiple options for assessing functional movement characteristics; however, many of these solutions have significant limitations. For example, wearable electromagnetic sensors are readily available [8], [9]; but data are affected by gravity noise and signal drift [10]. Additionally, this technology is costly and requires technical expertise for data analysis. Alternatively, the Kinect sensor is a commercially-available, cost-effective video game accessory [11], [12] capable of extracting data from 3D skeletal modeling [13]. The Kinect’s validity, and its use with various biomechanical applications, has been examined previously [14], [15], [16], [17]. While the first version of the Kinect (v1) had poor accuracy and tracking capacity [18], the new version (Kinect v2) may provide improved skeletal tracking with its higher camera and depth resolutions. The color (1920 × 1080 pixels) and depth (512 × 424 pixels) resolution of the Kinect v2 was significantly higher than the v1, allowing more precise joint trajectory tracking [2], [3], [4], [5]. The Kinect’s technological advancement, ease of data acquisition and processing software, increase its potential for accurately analyzing gait.

Poor agreement between Kinect v1 and optoelectronic motion capture systems for sagittal plane kinematic variables has been reported during treadmill lower extremity gait [19] and other functional movements [20]. These differences are not unexpected given the technological limitations of the Kinect v1. Studies have also demonstrated that the Kinect v1 and v2 are significantly better at assessing spatiotemporal parameters compared to lower extremity kinematic variables [21]. With the technological advancement of the v2 over the v1 camera, it is logical that tracking of sagittal plane joint range of motion during functional movement would significantly improve. But before the Kinect v2 can be utilized clinically for applications like gait analysis, its validity when assessing lower extremity kinematics must be established and compared to previous findings. To our knowledge, no studies have done this; therefore this study seeks to establish the validity of the Kinect v2 in assessing lower extremity sagittal plane kinematic and spatiotemporal parameters during treadmill gait analysis.

Section snippets

Subjects

Ten healthy subjects (5 males, 5 females, age: 26.7 ± 5.4 years, height: 174.4 ± 7.9 cm, mass: 71.8 ± 11.4 kg) participated in this study. All subjects were free of previous lower extremity surgery and were free of current injury that resulted in limitation of physical activity level. The study was approved by the University’s Human Subjects Review Board, and all participants provided written consent.

Subject preparation

Upon arrival at the laboratory, subjects were familiarized with the experimental setup and outfitted

Results

Participant characteristics are in Table 2.

Discussion

The study goal was to examine the capacity of the Kinect v2 to measure lower extremity sagittal plane kinematics and spatiotemporal events during treadmill walking. Our results indicate that the Kinect is an acceptable tool for the assessment of sagittal plane knee and hip ROM and joint angles across the gait cycle; however, significant limitations existed when assessing ankle joint kinematics. Additionally, the capacity of the Kinect to assess spatiotemporal variables was excellent for step

Conclusions

In conclusion, the Kinect v2 sensor can potentially to be an effective clinical tool for evaluating sagittal plane knee and hip joint kinematics and some spatiotemporal variables during walking gait. Given advances in sensor technology and ease of data acquisition and processing, the Kinect v2 may be a realistic and cost effective alternative to expensive traditional motion analysis systems for clinical applications. Additional investigations should examine the capacity of the Kinect v2 to

Conflicts of interest statement

None.

References (34)

  • C. Bunce

    Correlation, agreement, and Bland-Altman analysis: statistical analysis of method comparison studies

    Am. J. Ophthalmol.

    (2009)
  • C.C. Chen et al.

    Comparison of ICC and CCC for assessing agreement for data without and with replications

    Comput. Stat. Data Anal.

    (2008)
  • R.A. Clark et al.

    Concurrent validity of the Microsoft Kinect for assessment of spatiotemporal gait variables

    J. Biomech.

    (2013)
  • N.L. Keijsers et al.

    Ambulatory motor assessment in Parkinson's disease

    Mov. Disord.

    (2006)
  • J.M. Hausdorff et al.

    Dynamic markers of altered gait rhythm in amyotrophic lateral sclerosis

    J. Appl. Physiol. (1985)

    (2000)
  • D.M. Frost et al.

    Using the functional movement screen™ to evaluate the effectiveness of training

    J. Strength Cond. Res.

    (2012)
  • M.J. Faber et al.

    Clinimetric properties of the performance-oriented mobility assessment

    Phys. Ther.

    (2006)
  • Cited by (116)

    View all citing articles on Scopus
    View full text