Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2016 Feb 5.
Published in final edited form as: J Biomech. 2014 Dec 10;48(3):544–548. doi: 10.1016/j.jbiomech.2014.11.048

A low cost real-time motion tracking approach using webcam technology

Chandramouli Krishnan 1,✉, Edward P Washabaugh 1, Yogesh Seetharaman 1
PMCID: PMC4306621  NIHMSID: NIHMS648164  PMID: 25555306

Abstract

Physical therapy is an important component of gait recovery for individuals with locomotor dysfunction. There is a growing body of evidence that suggests that incorporating a motor learning task through visual feedback of movement trajectory is a useful approach to facilitate therapeutic outcomes. Visual feedback is typically provided by recording the subject’s limb movement patterns using a three-dimensional motion capture system and displaying it in real-time using customized software. However, this approach can seldom be used in the clinic because of the technical expertise required to operate this device and the cost involved in procuring a three-dimensional motion capture system. In this paper, we describe a low cost two-dimensional real-time motion tracking approach using a simple webcam and an image processing algorithm in LabVIEW Vision Assistant. We also evaluated the accuracy of this approach using a high precision robotic device (Lokomat) across various walking speeds. Further, the reliability and feasibility of real-time motion-tracking were evaluated in healthy human participants. The results indicated that the measurements from the webcam tracking approach were reliable and accurate. Experiments on human subjects also showed that participants could utilize the real-time kinematic feedback generated from this device to successfully perform a motor learning task while walking on a treadmill. These findings suggest that the webcam motion tracking approach is a feasible low cost solution to perform real-time movement analysis and training.

Keywords: Real time, Motion tracking, Motor learning, Gait, Kinematics

Introduction

Gait abnormalities are common in individuals with a variety of neurological and orthopedic disorders (Beltran et al., 2014; Kao et al., 2014; Lamontagne et al., 2002; Lewek et al., 2002; Sosnoff et al., 2012). Physical therapy plays an important role in facilitating the recovery of locomotion (Hesse, 2001; Richards et al., 1999). Accordingly, therapists often spend a considerable amount of time and effort assessing and retraining locomotion with their patients. When assessing movement patterns, therapists typically rely on qualitative evaluations to quantify gait abnormalities (e.g., visual observation or videographic analysis). Qualitative gait analysis requires minimal instrumentation and is fairly simple, inexpensive, and easier to implement in routine clinical practice (Burnfield and Norkin, 2013). However, considerable training and practice are necessary to learn the observational skills necessary to perform such qualitative analysis. Further, the outcomes are very subjective and are considered to lack sufficient reliability to make any meaningful conclusions from the analysis (Burnfield and Norkin, 2013; Krebs et al., 1985).

Quantitative gait analysis addresses these limitations, and is increasingly sought by third-party payers when assessing patient function, establishing therapeutic strategies, and documenting patient progress. While there are some low cost devices for quantifying spatiotemporal parameters of gait (e.g., cadence, step length and symmetry, step and stride duration, etc.) (Bergamini et al., 2013; Salarian et al., 2010; Spain et al., 2012), there are not many low cost instrumentations for kinematic evaluation (Bonnet et al., 2012; Picerno et al., 2008). Further, these devices typically don’t provide real-time data, which may be of particular interest to clinicians due to its application for gait training in virtual environment. An electrogoniometer is low cost and can provide real-time kinematic data (Peat et al., 1976); however, it can interfere with natural movement patterns due to cumbersome attachments and is less accurate when used for joints such as the hip or ankle (Manal and Buchanan, 2003). The Microsoft Kinect sensor device allows for economical and non-intrusive gait and motion analysis (Bonnechere et al., 2014; Clark et al., 2013; Clark et al., 2012; Galna et al., 2014; Schmitz et al., 2014). However, the software development kits are primarily designed to extract spatiotemporal parameters and typically don’t work well for sagittal plane kinematics and when joints lie close to each other (Bonnechere et al., 2012; Bonnechere et al., 2014; Pfister et al., 2014).

In this paper, we describe a novel low cost marker-based approach for real-time two-dimensional motion tracking using a webcam and an image processing algorithm in LabVIEW Vision Assistant. We also provide the results from a validation experiment performed using a high-precision lower-extremity gait robot called the Lokomat. Further, we report the feasibility of real-time motion tracking using an established target-tracking approach that has the potential to facilitate gait recovery after stroke.

Methods

Hardware and Algorithm for real-time tracking

The hardware required for acquiring the marker data were procured from Noraxon USA, Inc. (Scottsdale, AZ). These included a Logitech HD Pro Webcam C920 (1080p, 30 FPS), a SunPAK 6600DX heavy duty tripod, a Rigid Industries floodlight, and standard 19mm retroreflective markers. The camera was connected to a Windows computer via an USB 2.0 cable. All data were collected and processed using custom-written programs in LabVIEW and Vision Assistant, version 2011 (National Instruments Corp., Austin, TX, USA). The steps involved in processing the data are as follows:

First, the program captures a frame of the video using IMAQdx Grab VI [Video Mode = 32 (800×600 MJPJ 30FPS), Gain = 255 (maximum), Exposure = 0 (minimum); Brightness = 128 (medium); and Contrast = 1.28 (medium)]. A region of interest is then selected using the IMAQ ConstructROI VI. The program filters the image data for the brightest (whitest) parts of the image using the IMAQ ColorThreshold VI and Binary Inverse VI. Here, pixels past a certain whiteness threshold are set to 1, the rest are set to 0. Thus, all pixels containing markers are set to 1. Next, the program discards everything outside the region of interest using the IMAQ ROIToMask 2 VI. The program looks at pixel values of 1 and selects items that are in the shape of a circle within a specified radius using IMAQ Find Circles VI. Finally, the program outputs real-time pixel coordinates of all these circles (i.e., the markers).

Real-time 2D Kinematic Tracking

A three-point model can be created from the hip, knee, and ankle markers to obtain two-dimensional kinematics of the hip and the knee joint during walking using the following equations.

HipAngle=arctan2([xknee-xHip]∣[yHip-yknee])kneeAngle=(90-HipAngle)-(arctan2([yankle-yknee]∣[xankle-xknee]))

Where Hip (relative to the vertical trunk) and Knee Angles represent the anatomical joint angles, xhip, xknee and xankle represent the x-coordinates, and yhip, yknee and yankle represent the y-coordinates of the markers over the respective anatomical landmarks.

Validation experiment using the Lokomat Gait Robot

A validation experiment using a lower extremity driven gait orthoses called as the Lokomat was conducted to validate the kinematic data obtained using our webcam tracking methodology. The Lokomat is a robotic device commonly used for gait training in individuals with neurological disorders (Jezernik et al., 2003; Krishnan et al., 2013a; Mayr et al., 2007). The device has four linear actuators for controlling the hip and knee joint motions and four potentiometers to measure hip and knee joint angles. We used the Lokomat system as our test bed for validating the results from our tracking system for two reasons: (1) The device can be configured to a ‘position control mode’, where the robot’s stiffness is high, which enables it to impose predefined motions with high repeatability and precision (Jezernik et al., 2003) and (2) The Lokomat provides movements only in the sagittal plane, which eliminates the possibility of errors due to off-plane motions.

Miniature battery operated LED markers were placed over the hip joint, knee joint, and the distal end of the Lokomat’s leg for tracking robotic gait movements. We used miniature LED markers for motion tracking as passive reflective markers cannot be tracked well over reflective surfaces. The pelvis of the Lokomat was secured to the rails of the Lokomat treadmill system to minimize unwanted vertical oscillatory movements. The robotic legs of the Lokomat were then set to move on a predefined gait trajectory at several walking speeds (1.0, 1.2, 1.5, 1.7, 2.0, 2.2, 2.5, 2.7, and 3.0 km/h randomly ordered). Kinematic data were recorded simultaneously from the potentiometers of the Lokomat and from the Webcam system for 2 minutes at each testing speed.

Human Subjects Experiment for assessing reliability and feasibility of real-time tracking

Experimental data were collected from four young healthy adults on two consecutive days to test reliability and feasibility of real-time target tracking. Prior to participation, subjects signed an informed consent document approved by the University of Michigan Human Subjects Institutional Review Board. Three 19mm retroreflective markers were placed over the subject’s greater trochanter, lateral femoral epicondyle, and lateral malleolus. The subject then walked over a motorized treadmill (Woodway USA) with their hands placed on a custom built treadmill rail system (Figure 1A). The kinematic data during treadmill walking were captured using the real-time marker tracking algorithm described above. The baseline kinematic data collected for 1 minute during normal walking were then ensemble averaged across gait cycles and scaled to generate a target-template trajectory. The target-template trajectory corresponded to a gait pattern that required increasing the hip and knee joint angle by a scale of 30% during the swing phase of the gait and was projected in the end-point space (i.e., spatial path of subject’s lateral malleolus on the sagittal plane) (Figure 1B) (Krishnan et al., 2013b; Krishnan et al., 2012). A forward kinematic analysis on the hip and knee joint angles was used to obtain the position of the subject’s ankle lateral malleolus (xa, ya), relative to greater trochanter. The following equation was used for the forward kinematic analysis:

[xaya]=[sinθh-sin(θk-θh)-cosθh-cos(θk-θh)][l1l2]

where l1 is the distance between the markers over the greater trochanter and lateral femoral epicondyle, l2 is the distance between the markers over the lateral femoral epicondyle and ankle lateral malleolus, θh is the hip joint angle, and θk is the knee joint angle (Figure 1A).

Figure 1.

Figure 1

Open in a new tab

(A) Schematic representation of real-time tracking set-up. (B) Schematics of target-template and (C) tracking error computation.

The following equation was used to generate the desired target-template trajectory from the baseline kinematic data:

[xtargetytarget]=[xbayba]+0.3[xhwyhw]

where xhw and yhw represent the Hanning-windowed version of the baseline trajectories (xba, yba). The target-template was then displayed concurrently with the participant’s actual ankle trajectory on a computer monitor placed in front of the participant (Figure 1A). The participant was then instructed to match the target continuously for 1 minute by modifying the kinematics of their testing leg. The participant performed 10 blocks of target tracking with one minute of rest between each block. The testing was repeated on their second visit.

Data Analysis

The mean absolute differences in hip and knee joint angles between the values recorded from the potentiometers of the Lokomat and the marker-based kinematic tracking at various gait speeds were computed to calculate the magnitude of error in real-time motion tracking. The number of strides and the stride duration were also computed and compared at each testing speed to determine the validity of the spatiotemporal features extracted from the webcam tracking approach. The mean absolute differences in hip and knee joint angles recorded from human subjects during treadmill walking on different days were computed to assess the reliability of data obtained across testing days. The feasibility of utilizing real-time motion tracking to successfully assist in modifying the kinematics of the participant’s leg movements to match a desired target-template was evaluated by computing the changes in tracking error observed during the 10 blocks of target tracking (Figure 1C). The tracking errors were normalized to those observed during the first block of target tracking.

Results

The kinematic data obtained from the webcam tracking approach were similar to those recorded from the potentiometers of the Lokomat system (Supplementary Figure 1). The mean absolute error was less than 2° for all velocities tested in this study (Table 1). The number of strides and stride duration at each of the gait speeds were almost identical between those computed from the webcam tracking and the Lokomat system (Table 1). The data from human subjects indicated that the observed hip and knee kinematics profiles were similar to those that have been reported in the literature and were reproducible between days (Figure 2). The mean absolute differences in the hip and knee joint kinematic data recorded on the two days were 1.56° ± 0.13° and 2.56° ± 0.53°, respectively. Results from target-tracking experiment also indicated that participants were able to utilize the real-time kinematic feedback to modify their foot trajectory and accurately match the target-template projected on the screen (Figure 3A). The tracking error observed during each block of training reduced consistently in all participants and were retained when tested on Day 2 (Figure 3B).

Table 1.

Data showing differences in kinematic and spatiotemporal parameters recorded in real-time using the position encoders of the Lokomat robotic device and the marker-based kinematic tracking using the webcam at various gait speeds. Values are computed from a 2-minute block recorded at each of the gait speeds tested. Absolute error values are provided as mean ± S.D. in degrees. Stride duration values are reported as mean ± S.D. in seconds.

Speed Mean Absolute Error Strides Stride Duration (s)
Hip Angle Knee Angle Lokomat Webcam Lokomat Webcam
1.0 km/h 0.35° ± 0.21° 1.44° ± 0.25° 28 28 4.22 ± 0.001 4.22 ± 0.002
1.2 km/h 0.64° ± 0.16° 1.47° ± 0.36° 34 34 3.52 ± 0.001 3.52 ± 0.002
1.5 km/h 0.21° ± 0.10° 1.43° ± 0.35° 43 43 2.81 ± 0.001 2.81 ± 0.001
1.7 km/h 0.65° ± 0.18° 1.49° ± 0.44° 49 49 2.48 ± 0.001 2.48 ± 0.000
2.0 km/h 0.24° ± 0.12° 1.47° ± 0.49° 57 57 2.11 ± 0.001 2.11 ± 0.001
2.2 km/h 0.69° ± 0.19° 1.47° ± 0.53° 63 63 1.92 ± 0.001 1.92 ± 0.001
2.5 km/h 0.73° ± 0.27° 1.49° ± 0.67° 71 71 1.69 ± 0.001 1.69 ± 0.001
2.7 km/h 0.71° ± 0.25° 1.53° ± 0.67° 76 76 1.56 ± 0.001 1.56 ± 0.001
3.0 km/h 0.66° ± 0.27° 1.64° ± 0.77° 85 85 1.41 ± 0.001 1.41 ± 0.001
Open in a new tab

Figure 2.

Figure 2

Open in a new tab

Hip and knee joint kinematics recorded using the webcam tracking approach from four subjects on two consecutive days.

Figure 3.

Figure 3

Open in a new tab

(A) Example of target-tracking performance from a subject at the beginning (block 1) and end of target-tracking practice (block 10). (B) Mean target-tracking error from four subjects during each block of training on two consecutive days. Target-tracking error is shown as % error observed during block 1 of practice on Day 1. Dotted vertical line separates performance on Day 1 and Day 2. Error bars represent standard error of the mean.

Discussion

There are many affordable solutions for offline human movement analyses. However, there aren’t many options for real-time motion tracking using low cost technology. This study addresses this gap by providing a novel measurement tool for real-time motion tracking using a commercially available webcam (Logitech C920). Our results from validation experiments using a high precision robotic device (Lokomat) and healthy human participants indicated that the measurements obtained are reliable and accurate. Experiments on human subjects also showed that participants could successfully utilize the feedback generated from this device to modify their gait patterns. These results show potential for the device as a low cost substitute for motion tracking and gait therapy that could be utilized in a clinical setting.

The costs associated with motion tracking device used in this study are for the webcam($70), hardware (tripod, flood light, markers, USB cable) (~$300), computer, and NI Vision Assistant runtime engine (~$730). After procuring the necessary hardware and software, using a free NI LabVIEW Run-Time engine, interested users with access to the custom LabVIEW executable file can then run the tests described in this study. Further, the real-time foot trajectory tracking approach described earlier can be easily modified to perform upper extremity reaching motions targeting reaching workspace (Supplementary Figure 2) (Ellis et al., 2007; Ellis et al., 2011), thereby serving as a potential low cost therapeutic tool for upper extremity rehabilitation. We expect that this application will not only benefit clinicians, but also potentially benefit researchers in developing countries as they typically cannot afford high-end motion capture systems.

There are some potential limitations to the described methodology. The approach described effectively provides real time analysis and therapy for a single plane of motion; however, it is to be recognized that out of plane motion would cause errors in estimated joint angles (Nielsen and Daugaard, 2008). Further, care should be taken to avoid marker occlusions as the current algorithm is not capable of handling missing markers, particularly because of a single-camera set-up. Finally, the use of this device needs to be restricted to tracking unilateral movements encountered at normal gait speeds, as commercially available webcams (including the one used in this study) are typically limited to 30 fps. It is important to note that a conventional three-dimensional system could account for all these issues and should be the primary choice if cost/time is not an issue.

Supplementary Material

1. Supplementary Figure 1.

(A) Comparison of cycle-by-cycle hip joint kinematics recorded from the Lokomat system and the webcam tracking approach at various gait speeds. The black traces represent joint angles recorded from the Lokomat system and the gray (or red) traces represent joint angles recorded from the webcam. For the purpose of clarity, only 5 of the 9 gait speeds tested are shown in the figure. (B) Comparison of cycle-by-cycle knee joint kinematics recorded from the Lokomat system and the webcam tracking approach at various gait speeds. The black traces represent joint angles recorded from the Lokomat system and the gray (or red) traces represent joint angles recorded from the webcam. For the purpose of clarity, only 5 of the 9 gait speeds tested are shown in the figure.

2. Supplementary Figure 2.

(A) Example of reaching workspace recorded from a single subject during planar reaching movements. The gray traces represent cycle-by-cycle reaching kinematics and the solid black trace represents the ensemble average of reaching kinematics. (B) Schematics of target-template construction (Target) from the ensemble average of reaching kinematics (Actual) to demonstrate the potential of using webcam tracking approach to improve reaching workspace. Note that the X- and Y-axes units are in meters [m].

Acknowledgments

This work was supported in part by funding from the University of Michigan’s Rehabilitation Robotics Interdisciplinary Faculty Initiative Program and from the National Institutes of Health (NIH R01 EB019834) to Dr. Krishnan.

Footnotes

Conflict of Interest Statement

None of the authors received any significant financial support for this study that could have influenced its outcome. The authors declare no conflicts of interest.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Beltran EJ, Dingwell JB, Wilken JM. Margins of stability in young adults with traumatic transtibial amputation walking in destabilizing environments. J Biomech. 2014;47:1138–1143. doi: 10.1016/j.jbiomech.2013.12.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Bergamini E, Guillon P, Camomilla V, Pillet H, Skalli W, Cappozzo A. Trunk inclination estimate during the sprint start using an inertial measurement unit: a validation study. J Appl Biomech. 2013;29:622–627. doi: 10.1123/jab.29.5.622. [DOI] [PubMed] [Google Scholar]
  3. Bonnechere B, Jansen B, Salvia P, Bouzahouene H, Omelina L, Cornelis J, Rooze M, Van Sint Jan S. Year What are the current limits of the Kinect sensor. Proc 9th Intl Conf. Disability, Virutal Reality & Associated Technologies; Laval, France.. [Google Scholar]
  4. Bonnechere B, Jansen B, Salvia P, Bouzahouene H, Omelina L, Moiseev F, Sholukha V, Cornelis J, Rooze M, Van Sint Jan S. Validity and reliability of the Kinect within functional assessment activities: comparison with standard stereophotogrammetry. Gait Posture. 2014;39:593–598. doi: 10.1016/j.gaitpost.2013.09.018. [DOI] [PubMed] [Google Scholar]
  5. Bonnet V, Mazza C, Fraisse P, Cappozzo A. A least-squares identification algorithm for estimating squat exercise mechanics using a single inertial measurement unit. J Biomech. 2012;45:1472–1477. doi: 10.1016/j.jbiomech.2012.02.014. [DOI] [PubMed] [Google Scholar]
  6. Burnfield J, Norkin CC. Examination of Gait. In: O’Sullivan SB, Schmitz TJ, Fulk GD, editors. Physical Rehabilitation. 6. FA Davis Company; 2013. [Google Scholar]
  7. Clark RA, Pua YH, Bryant AL, Hunt MA. Validity of the Microsoft Kinect for providing lateral trunk lean feedback during gait retraining. Gait Posture. 2013;38:1064–1066. doi: 10.1016/j.gaitpost.2013.03.029. [DOI] [PubMed] [Google Scholar]
  8. Clark RA, Pua YH, Fortin K, Ritchie C, Webster KE, Denehy L, Bryant AL. Validity of the Microsoft Kinect for assessment of postural control. Gait Posture. 2012;36:372–377. doi: 10.1016/j.gaitpost.2012.03.033. [DOI] [PubMed] [Google Scholar]
  9. Ellis MD, Acosta AM, Yao J, Dewald JP. Position-dependent torque coupling and associated muscle activation in the hemiparetic upper extremity. Exp Brain Res. 2007;176:594–602. doi: 10.1007/s00221-006-0637-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Ellis MD, Kottink AI, Prange GB, Rietman JS, Buurke JH, Dewald JP. Quantifying loss of independent joint control in acute stroke with a robotic evaluation of reaching workspace. Conf Proc IEEE Eng Med Biol Soc. 2011;2011:8231–8234. doi: 10.1109/IEMBS.2011.6091940. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Galna B, Barry G, Jackson D, Mhiripiri D, Olivier P, Rochester L. Accuracy of the Microsoft Kinect sensor for measuring movement in people with Parkinson’s disease. Gait Posture. 2014;39:1062–1068. doi: 10.1016/j.gaitpost.2014.01.008. [DOI] [PubMed] [Google Scholar]
  12. Hesse S. Locomotor therapy in neurorehabilitation. NeuroRehabilitation. 2001;16:133–139. [PubMed] [Google Scholar]
  13. Jezernik S, Colombo G, Keller T, Frueh H, Morari M. Robotic orthosis lokomat: a rehabilitation and research tool. Neuromodulation. 2003;6:108–115. doi: 10.1046/j.1525-1403.2003.03017.x. [DOI] [PubMed] [Google Scholar]
  14. Kao PC, Dingwell JB, Higginson JS, Binder-Macleod S. Dynamic instability during post-stroke hemiparetic walking. Gait Posture. 2014;40:457–463. doi: 10.1016/j.gaitpost.2014.05.014. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Krebs DE, Edelstein JE, Fishman S. Reliability of observational kinematic gait analysis. Phys Ther. 1985;65:1027–1033. doi: 10.1093/ptj/65.7.1027. [DOI] [PubMed] [Google Scholar]
  16. Krishnan C, Kotsapouikis D, Dhaher YY, Rymer WZ. Reducing robotic guidance during robot-assisted gait training improves gait function: a case report on a stroke survivor. Arch Phys Med Rehabil. 2013a;94:1202–1206. doi: 10.1016/j.apmr.2012.11.016. [DOI] [PubMed] [Google Scholar]
  17. Krishnan C, Ranganathan R, Dhaher YY, Rymer WZ. A pilot study on the feasibility of robot-aided leg motor training to facilitate active participation. PLoS One. 2013b;8:e77370. doi: 10.1371/journal.pone.0077370. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Krishnan C, Ranganathan R, Kantak SS, Dhaher YY, Rymer WZ. Active robotic training improves locomotor function in a stroke survivor. J Neuroeng Rehabil. 2012;9:57. doi: 10.1186/1743-0003-9-57. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Lamontagne A, Malouin F, Richards CL, Dumas F. Mechanisms of disturbed motor control in ankle weakness during gait after stroke. Gait Posture. 2002;15:244–255. doi: 10.1016/s0966-6362(01)00190-4. [DOI] [PubMed] [Google Scholar]
  20. Lewek M, Rudolph K, Axe M, Snyder-Mackler L. The effect of insufficient quadriceps strength on gait after anterior cruciate ligament reconstruction. Clin Biomech (Bristol, Avon) 2002;17:56–63. doi: 10.1016/s0268-0033(01)00097-3. [DOI] [PubMed] [Google Scholar]
  21. Manal KT, Buchanan TS. Biomechanics of human movement. In: Kutz M, editor. Standard handbook of biomedical engineering & design. The McGraw-Hill Companies, Inc; USA: 2003. [Google Scholar]
  22. Mayr A, Kofler M, Quirbach E, Matzak H, Frohlich K, Saltuari L. Prospective, blinded, randomized crossover study of gait rehabilitation in stroke patients using the Lokomat gait orthosis. Neurorehabil Neural Repair. 2007;21:307–314. doi: 10.1177/1545968307300697. [DOI] [PubMed] [Google Scholar]
  23. Nielsen DB, Daugaard M. Doctoral. Jönköping University; Jönköping, Sweden: 2008. Comparison of angular measurements by 2D and 3D gait analysis. [Google Scholar]
  24. Peat M, Graham RE, Fulford R, Quanbury AO. An electrogoniometer for the measurement of single plane movements. J Biomech. 1976;9:423–424. doi: 10.1016/0021-9290(76)90121-4. [DOI] [PubMed] [Google Scholar]
  25. Pfister A, West AM, Bronner S, Noah JA. Comparative abilities of Microsoft Kinect and Vicon 3D motion capture for gait analysis. J Med Eng Technol. 2014:1–7. doi: 10.3109/03091902.2014.909540. [DOI] [PubMed] [Google Scholar]
  26. Picerno P, Cereatti A, Cappozzo A. Joint kinematics estimate using wearable inertial and magnetic sensing modules. Gait Posture. 2008;28:588–595. doi: 10.1016/j.gaitpost.2008.04.003. [DOI] [PubMed] [Google Scholar]
  27. Richards CL, Malouin F, Dean C. Gait in stroke: assessment and rehabilitation. Clin Geriatr Med. 1999;15:833–855. [PubMed] [Google Scholar]
  28. Salarian A, Horak FB, Zampieri C, Carlson-Kuhta P, Nutt JG, Aminian K. iTUG, a sensitive and reliable measure of mobility. IEEE Trans Neural Syst Rehabil Eng. 2010;18:303–310. doi: 10.1109/TNSRE.2010.2047606. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Schmitz A, Ye M, Shapiro R, Yang R, Noehren B. Accuracy and repeatability of joint angles measured using a single camera markerless motion capture system. J Biomech. 2014;47:587–591. doi: 10.1016/j.jbiomech.2013.11.031. [DOI] [PubMed] [Google Scholar]
  30. Sosnoff JJ, Sandroff BM, Motl RW. Quantifying gait abnormalities in persons with multiple sclerosis with minimal disability. Gait Posture. 2012;36:154–156. doi: 10.1016/j.gaitpost.2011.11.027. [DOI] [PubMed] [Google Scholar]
  31. Spain RI, St George RJ, Salarian A, Mancini M, Wagner JM, Horak FB, Bourdette D. Body-worn motion sensors detect balance and gait deficits in people with multiple sclerosis who have normal walking speed. Gait Posture. 2012;35:573–578. doi: 10.1016/j.gaitpost.2011.11.026. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1. Supplementary Figure 1.

(A) Comparison of cycle-by-cycle hip joint kinematics recorded from the Lokomat system and the webcam tracking approach at various gait speeds. The black traces represent joint angles recorded from the Lokomat system and the gray (or red) traces represent joint angles recorded from the webcam. For the purpose of clarity, only 5 of the 9 gait speeds tested are shown in the figure. (B) Comparison of cycle-by-cycle knee joint kinematics recorded from the Lokomat system and the webcam tracking approach at various gait speeds. The black traces represent joint angles recorded from the Lokomat system and the gray (or red) traces represent joint angles recorded from the webcam. For the purpose of clarity, only 5 of the 9 gait speeds tested are shown in the figure.

NIHMS648164-supplement-1.jpg (469.8KB, jpg)
2. Supplementary Figure 2.

(A) Example of reaching workspace recorded from a single subject during planar reaching movements. The gray traces represent cycle-by-cycle reaching kinematics and the solid black trace represents the ensemble average of reaching kinematics. (B) Schematics of target-template construction (Target) from the ensemble average of reaching kinematics (Actual) to demonstrate the potential of using webcam tracking approach to improve reaching workspace. Note that the X- and Y-axes units are in meters [m].

NIHMS648164-supplement-2.JPG (583.7KB, JPG)

RESOURCES