No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Accuracy and repeatability of joint angles measured using a single camera markerless motion capture system
... Compared to the marker-based motion capture system, advantages of the markerless motion capture system include eliminating variations in marker placements to reduce inter-session measurement variability, reducing patient/subject preparation time, and cost-effectiveness. [9][10][11][12][13][14][15][16][17][18][19][20] Previous studies on markerless motion capture systems have focused on gait or lower extremity movements. 12,14,16,21 Several recent studies have also proposed a computer vision-based markerless motion capture system for assessing upper extremity and hand motor function in people with neurologic disorders. ...
... [9][10][11][12][13][14][15][16][17][18][19][20] Previous studies on markerless motion capture systems have focused on gait or lower extremity movements. 12,14,16,21 Several recent studies have also proposed a computer vision-based markerless motion capture system for assessing upper extremity and hand motor function in people with neurologic disorders. [22][23][24][25] These markerless motion capture systems mainly used 3-dimensional (3D) depth-sensing cameras, such as the Kinect, and skeleton models to quantify joint angles. ...
... Using a gold-standard marker-based motion capture system, researchers have shown a high level of accuracy, validity, and reliability for various markerless motion capture systems to record human body kinematics. 12,14,15,21,26,27 In addition to recording human kinematics, the computer vision-based markerless motion capture system can be used to detect and track objects that a person manipulates. Using this approach, we can also assess improvement in motor performance or learning of ne hand motor skills, speci cally related to object manipulation. ...
Computer vision-based 3D motion capture systems are useful for human movement kinematic evaluation. We developed a computer vision-based 3D motion capture system by tracking an object manipulated by a hand to assess fine hand motor function. This study aimed to determine the accuracy and feasibility of this system to assess fine hand motor skill learning. We conducted four different experiments to test the accuracy and feasibility of this system. We used two action cameras with high resolution and high frame rate. We tested our system's accuracy in estimating the 3D positions of a static object and a moving object in different directions. We tested the feasibility of this system to assess fine hand motor skill learning in four non-disabled young adults. This clinical feasibility experiment used a standard motor skill learning experimental design. We utilized color-based object detection and tracking. Our results support that this computer vision-based motion capture system is accurate in reconstructing the 3D positions of a moving object. Also, we demonstrated that the system can be used to quantify the kinematics of the target object movements to evaluate fine hand motor skill learning. Future studies are needed to establish the reliability and validity of this system to assess fine hand motor skills in patient populations.
... There have been several previous studies comparing motion capture systems that have focused on one-to-one comparisons of a single test system and a gold standard system [29,32,[35][36][37][38][39][40][41], studied the lower limbs [26,30,40,[42][43][44][45], or relied on mechanical testing devices to ensure the greatest replicability of the ground truth [36,[46][47][48]. For the oneto-one system comparisons, the parameters examined, motions selected, and populations tested varied greatly, rendering cross-system conclusions impractical. ...
... The Vicon optical marker-based system was selected as the reference system based on its popularity and usage in the literature [27,[36][37][38]44,45,48,[65][66][67][68]. The IMU-based [28][29][30][31][32][39][40][41][42][43][69][70][71] and markerless systems [13,[22][23][24][25]27,38,[44][45][46][47]72,73] were selected due to popularity in the literature and due to their differing mechanisms of motion capture. Please see Appendix B for diagrams of the sensor placements for the systems and the camera placements. ...
... The markerless shoulder bias values ( Figure 3B) measured in this study were also notably different from those found in the literature (current study measured approximately −25 • compared to an average around +10 • ) [26,27]. Although this difference is large, it may be a more accurate representation of the expected performance of these motion capture systems given the use of complex tasks [76] and human subjects in this study compared to simple ROM measurements [26,27] and testing machines [47] found in the literature. ...
Current literature lacks a comparative analysis of different motion capture systems for tracking upper limb (UL) movement as individuals perform standard tasks. To better understand the performance of various motion capture systems in quantifying UL movement in the prosthesis user population, this study compares joint angles derived from three systems that vary in cost and motion capture mechanisms: a marker-based system (Vicon), an inertial measurement unit system (Xsens), and a markerless system (Kinect). Ten healthy participants (5F/5M; 29.6 ± 7.1 years) were trained with a TouchBionic i-Limb Ultra myoelectric terminal device mounted on a bypass prosthetic device. Participants were simultaneously recorded with all systems as they performed standardized tasks. Root mean square error and bias values for degrees of freedom in the right elbow, shoulder, neck, and torso were calculated. The IMU system yielded more accurate kinematics for shoulder, neck, and torso angles while the markerless system performed better for the elbow angles. By evaluating the ability of each system to capture kinematic changes of simulated upper limb prosthesis users during a variety of standardized tasks, this study provides insight into the advantages and limitations of using different motion capture technologies for upper limb functional assessment.
... However, inherent limits in data collecting may restrict its use in contexts such as patient homes, sports fields, or public spaces where the use of a large number of cameras is impractical. Here, a markerless motion capture system has been offered as one possible solution [1,2]. ...
... Thus, investigations examining the validity of such systems under different conditions are crucial. In this context, doctors, sports practitioners, and researchers have been paying close attention to a markerless motion capture system [1,[3][4][5][6][7]. ...
... The validity of the Kinect™ sensor, created first for interacting with video games on the Microsoft Xbox™ platform by using body movements, for the analysis of gait parameters has been previously evaluated [1,[3][4][5][6][7]. Various pieces of software, including various filters and calibrations, have been studied in these works. ...
The use of markerless motion capture systems is becoming more popular for walking and running analysis given their user-friendliness and their time efficiency but in some cases their validity is uncertain. Here, the test-retest reliability of the MotionMetrix software combined with the use of Kinect sensors is tested with 24 healthy volunteers for walking (at 5 km•h −1) and running (at 10 and 15 km•h −1) gait analysis in two different trials. All the parameters given by the MotionMetrix software for both walking and running gait analysis are tested in terms of reliability. No significant differences (p > 0.05) were found for walking gait parameters between both trials except for the phases of loading response and double support, and the spatiotemporal parameters of step length and step frequency. Additionally, all the parameters exhibit acceptable reliability (CV < 10%) but step width (CV > 10%). When analyzing running gait, although the parameters here tested exhibited different reliability values at 10 km•h −1 , the system provided reliable measurements for most of the kinematic and kinetic parameters (CV < 10%) when running at 15 km•h −1. Overall, the results obtained show that, although some variables must be interpreted with caution, the Kinect + Motion-Metrix system may be useful for walking and running gait analysis. Nevertheless, the validity still needs to be determined against a gold standard system to fully trust this technology and software combination.
... Additionally, SHRED far outperforms the linear and SDN architectures. While more input sensors generally increases reconstruction accuracy, performance with a mean-squared error as little as 0.068 and 0.048 degrees (rotational variables) can be achieved with one and three sensors with SHRED, well below the accuracy and repeatability recommended even for clinical gait analyses (2 deg in sagittal-plane and 5 deg in frontal-plane) [48], [50]. Table 2 provides a summary of the mean-squared errors for the rotational kinematic variables across modeling architectures and conditions. ...
... Four combinations of dynamic trajectory inputs are tested: (1) three randomly chosen kinematic features (transverse-plane pelvis rotation angle, medio-lateral pelvis position, right hip adduction angle), (2) three nonrandomly chosen kinematic features (right hip flexion angle, right knee flexion angle, right ankle dorsiflexion angle), (3) one randomly chosen kinematic feature (medio-lateral pelvis position), and (4) one non-randomly chosen kinematic feature (right ankle dorsiflexion angle). The non-randomly chosen features represent the kinematics most commonly used for biomechanical assessments, and which generally have the greatest accuracy and inter-session reliability (i.e, sagittalplane kinematics) [50]. Fig. 4 shows the performance of SHRED relative to the SDN and linear models for reconstructing human motion across individuals. ...
Sensing is one of the most fundamental tasks for the monitoring, forecasting, and control of complex, spatio-temporal systems. In many applications, a limited number of sensors are mobile and move with the dynamics, with examples including wearable technology, ocean monitoring buoys, and weather balloons. In these dynamic systems (without regions of statistical-independence), the measurement time history encodes a significant amount of information that can be extracted for critical tasks. Most model-free sensing paradigms aim to map current sparse sensor measurements to the high-dimensional state space, ignoring the time-history all together. Using modern deep learning architectures, we show that a sequence-to-vector model, such as an LSTM (long, short-term memory) network, with a decoder network, dynamic trajectory information can be mapped to full state-space estimates. Indeed, we demonstrate that by leveraging mobile sensor trajectories with shallow recurrent decoder networks, we can train the network (i) to accurately reconstruct the full state space using arbitrary dynamical trajectories of the sensors, (ii) the architecture reduces the variance of the mean-squared error of the reconstruction error in comparison with immobile sensors, and (iii) the architecture also allows for rapid generalization (parameterization of dynamics) for data outside the training set. Moreover, the path of the sensor can be chosen arbitrarily, provided training data for the spatial trajectory of the sensor is available. The exceptional performance of the network architecture is demonstrated on three applications: turbulent flows, global sea-surface temperature data, and human movement biomechanics.
... Any error in the measured marker position will thus affect the subsequent calculations and ultimately the conclusions drawn from the motion capture data. If other sensors like 2D cameras and inertial measurement units are used to measure human motion, these are often benchmarked against values calculated from 3D optical motion capture data [1][2][3][4][5], without considering that the data from 3D Table 1 Summary of the error sources affecting the markers, alongside the literature averages (lit. avg., summarised from [11][12][13][14]) and assumed errors that are used for the simulation in this work with the maximum modelled deviation ∑ . ...
... For step (1), fuzzy uncertainty could be considered, but we assume these measures to be deterministic since these quantities are usually measured manually or involve a fitting or optimisation process. Similarly, for step (2), fuzzy uncertainty could be used, but we assume that the initial construction is accurate and that soft tissue artefacts are unlikely due to the static measurement. ...
Measurement uncertainty is present when using optical motion capture to investigate joint angles during human gait. The aim of this work is to analyse the effect of uncertainty in the position of optical markers on the calculation of joint angles during human gait.
The uncertainty is based on known errors of the measurement of marker positions (e. g. measurement inaccuracies, marker placement errors and soft tissue artefacts) and is modelled with epistemic uncertainty. Based on optical motion capture data of normal human gait, the epistemic uncertainty is propagated through the joint angle calculations using the Graph Follower algorithm. While being well suited to model the uncertainty present in motion capture, epistemic uncertainty has not yet been considered to model errors in the marker positions.
The main contribution of this work is a new model for marker position errors and a new method to analyse the effects of these errors on the subsequent joint angle calculations during gait. This allows for an efficient worst case analysis, based on a real dataset, providing information on the largest possible error in the joint angle calculation based on given marker position errors. It is also capable of providing the opposite information, i. e. a given limit on joint angle errors can be used to determine acceptable marker errors. Further, it is also possible to investigate the effect of specific markers and their contribution to the joint angle analysis.
... Within a recent SWOT (i.e., strength, weakness, opportunity, and threat) analysis looking at portable and low-cost markerless motion capture systems, Armitano-Lago et al. [2] proposed that markerless motion capture systems show considerable promise with regard to enhancing our understanding of human movement characteristics, especially in providing unrestricted and simple movement assessments in natural sporting contexts. While still limited, a growing body of literature has proposed the validity of markerless motion capture systems when compared to marker-based systems [11][12][13][14][15]. For instance, Sandau et al. [11] suggested that a markerless motion capture system was able to reliably produce data within the sagittal and frontal plane of motion during walking (e.g., joint flexion, extension, abduction, and adduction). ...
... Looking at sagittal plane kinematics in a vertical jump task, Drazan et al. [12] found a very strong agreement between a custom markerless model approach and a gold-standard marker-based system. Further, Schmitz et al. [15] reported small differences in accuracy and reliability between a marker-based system and a single-camera markerless motion capture system. On the other hand, Harsted et al. [16] found reliability scores that were moderately acceptable for most measures but unacceptable for knee valgus and varus when comparing a markerless motion capture system to a traditional marker-based system during jumping tasks in preschool children. ...
With advancements in technology able to quantify wide-ranging features of human movement, the aim of the present study was to investigate the inter-device technological reliability of a three-dimensional markerless motion capture system (3D-MCS), quantifying different movement tasks. A total of 20 healthy individuals performed a test battery consisting of 29 different movements, from which 214 different metrics were derived. Two 3D-MCS located in close proximity were utilized to quantify movement characteristics. Independent sample t-tests with selected reliability statistics (i.e., intraclass correlation coefficient (ICC), effect sizes, and mean absolute differences) were used to evaluate the agreement between the two systems. The study results suggested that 95.7% of all metrics analyzed revealed negligible or small between-device effect sizes. Further, 91.6% of all metrics analyzed showed moderate or better agreement when looking at the ICC values, while 32.2% of all metrics showed excellent agreement. For metrics measuring joint angles (198 metrics), the mean difference between systems was 2.9 degrees, while for metrics investigating distance measures (16 metrics; e.g., center of mass depth), the mean difference between systems was 0.62 cm. Caution is advised when trying to generalize the study findings beyond the specific technology and software used in this investigation. Given the technological reliability reported in this study, as well as the logistical and time-related limitations associated with marker-based motion capture systems, it may be suggested that 3D-MCS present practitioners with an opportunity to reliably and efficiently measure the movement characteristics of patients and athletes. This has implications for monitoring the health/performance of a broad range of populations.
... Nonetheless, Markerless systems do not require markers to be placed on the participant, instead using synchronized 2D video cameras to achieve a 3D reconstruction. Markerless MCS have been developed as an alternative solution to overcome the limitations of marker-based MCS [5][6][7][8][9]. These systems eliminate the need for markers, providing a time-efficient and user-friendly option for capturing human motion data. ...
... These systems eliminate the need for markers, providing a time-efficient and user-friendly option for capturing human motion data. Unlike marker-based MCS, markerless systems rely on image-based tracking and machine-learning algorithms to estimate the movement of body joints and segments [5][6][7][8][9]. This provides a more natural experience for the subject being recorded, as well as a more practical and less time-consuming setting-up process for the practitioner. ...
Citation: Jaén-Carrillo, D.; García-Pinillos, F.; Chicano-Gutiérrez, J.M.; Pérez-Castilla, A.; Soto-Hermoso, V.; Molina-Molina, A.; Ruiz-Alias, S.A. Abstract: Markerless motion capture systems (MCS) have been developed as an alternative solution to overcome the limitations of 3D MCS as they provide a more practical and efficient setup process given, among other factors, the lack of sensors attached to the body. However, this might affect the accuracy of the measures recorded. Thus, this study is aimed at evaluating the level of agreement between a markerless MSC (i.e., MotionMetrix) and an optoelectronic MCS (i.e., Qualisys). For such purpose, 24 healthy young adults were assessed for walking (at 5 km/h) and running (at 10 and 15 km/h) in a single session. The parameters obtained from MotionMetrix and Qualisys were tested in terms of level of agreement. When walking at 5 km/h, the MotionMetrix system significantly underestimated the stance and swing phases, as well as the load and pre-swing phases (p < 0.05) reporting also relatively low systematic bias (i.e., ≤ −0.03 s) and standard error of the estimate (SEE) (i.e., ≤0.02 s). The level of agreement between measurements was perfect (r > 0.9) for step length left and cadence and very large (r > 0.7) for step time left, gait cycle, and stride length. Regarding running at 10 km/h, bias and SEE analysis revealed significant differences for most of the variables except for stride time, rate and length, swing knee flexion for both legs, and thigh flexion left. The level of agreement between measurements was very large (r > 0.7) for stride time and rate, stride length, and vertical displacement. At 15 km/h, bias and SEE revealed significant differences for vertical displacement, landing knee flexion for both legs, stance knee flexion left, thigh flexion, and extension for both legs. The level of agreement between measurements in running at 15 km/h was almost perfect (r > 0.9) when comparing Qualisys and MotionMetrix parameters for stride time and rate, and stride length. The agreement between the two motion capture systems varied for different variables and speeds of locomotion, with some variables demonstrating high agreement while others showed poor agreement. Nonetheless, the findings presented here suggest that the MotionMetrix system is a promising option for sports practitioners and clinicians interested in measuring gait variables, particularly in the contexts examined in the study.
... They concluded that Kinect is a reliable markerless tool that is suitable for use as a fast estimator of morphology. Schmitz et al. [5] validated the accuracy of Kinect in measuring knee joint angle of a jig by comparing its measurement using a digital inclinometer that acted as a ground-truth, and they reported that the performance of the Kinect system was satisfactory in terms of knee flexion and abduction. The accuracy of using a smartphone as a measurement system for joint angle has been reviewed by Mourcou et al. [6], who concluded that smartphone applications are reliable for clinical measurements of joint position and range of motion (ROM). ...
... Earlier in 2006, Mündermann et al. [7] described several methods of MMC video processing modules including background separation, visual hull, and iterative closest point methods, etc., and pointed out that MMC has the potential to achieve a level of accuracy that facilitates the biomechanics research of normal and pathological human movement. Together with the reliable performance of MMC technology in the measurement of joint angle and body movement as reflected by [5,6], it is suggested that the MMC system can be further applied to the rehabilitation field to measure patients' motor function. However, the actual application of MMC technology for clinical measurement in rehabilitation is still at a preliminary stage. ...
Background
Markerless motion capture (MMC) technology has been developed to avoid the need for body marker placement during motion tracking and analysis of human movement. Although researchers have long proposed the use of MMC technology in clinical measurement—identification and measurement of movement kinematics in a clinical population, its actual application is still in its preliminary stages. The benefits of MMC technology are also inconclusive with regard to its use in assessing patients’ conditions. In this review we put a minor focus on the method’s engineering components and sought primarily to determine the current application of MMC as a clinical measurement tool in rehabilitation.
Methods
A systematic computerized literature search was conducted in PubMed, Medline, CINAHL, CENTRAL, EMBASE, and IEEE. The search keywords used in each database were “Markerless Motion Capture OR Motion Capture OR Motion Capture Technology OR Markerless Motion Capture Technology OR Computer Vision OR Video-based OR Pose Estimation AND Assessment OR Clinical Assessment OR Clinical Measurement OR Assess.” Only peer-reviewed articles that applied MMC technology for clinical measurement were included. The last search took place on March 6, 2023. Details regarding the application of MMC technology for different types of patients and body parts, as well as the assessment results, were summarized.
Results
A total of 65 studies were included. The MMC systems used for measurement were most frequently used to identify symptoms or to detect differences in movement patterns between disease populations and their healthy counterparts. Patients with Parkinson’s disease (PD) who demonstrated obvious and well-defined physical signs were the largest patient group to which MMC assessment had been applied. Microsoft Kinect was the most frequently used MMC system, although there was a recent trend of motion analysis using video captured with a smartphone camera.
Conclusions
This review explored the current uses of MMC technology for clinical measurement. MMC technology has the potential to be used as an assessment tool as well as to assist in the detection and identification of symptoms, which might further contribute to the use of an artificial intelligence method for early screening for diseases. Further studies are warranted to develop and integrate MMC system in a platform that can be user-friendly and accurately analyzed by clinicians to extend the use of MMC technology in the disease populations.
... The benefits of the markerless motion capture system over the marker-based motion capture system include removing differences in marker placement to minimize inter-session measurement variability, minimizing patient/subject preparation time, and cost-effectiveness 7-18 . Previous studies on markerless motion capture systems concentrated on gait or lower extremity movements 11,13,15,19 . Several recent studies have also proposed a computer vision-based markerless motion capture approach for measuring upper extremity and hand motor function in people with neurologic disorders 20-23 . ...
... These markerless motion capture systems primarily utilized 3-dimensional (3D) depth-sensing cameras, such as the Kinect, to estimate joint kinematics. Researchers demonstrated a high degree of accuracy, validity, and reliability for various markerless motion capture systems to record human body kinematics by using a gold-standard marker-based motion capture system 11,13,14,19,24,25 . ...
We developed a computer vision-based three-dimension (3D) motion capture system employing two action cameras to examine fine hand motor skill by tracking an object manipulated by a hand. This study aimed to examine the accuracy and feasibility of this approach for detecting changes in a fine hand motor skill. We conducted three distinct experiments to assess the system's accuracy and feasibility. We employed two high-resolution, high-frame-rate action cameras. We evaluated the accuracy of our system in calculating the 3D locations of moving object in various directions. We also examined the system's feasibility in identifying improvement in fine hand motor skill after practice in eleven non-disabled young adults. We utilized color-based object detection and tracking to estimate the object's 3D location, and then we computed the object's kinematics, representing the endpoint goal-directed arm reaching movement. Compared to ground truth measurements, the findings demonstrated that our system can adequately estimate the 3D locations of a moving object. We also showed that the system can be used to measure the endpoint kinematics of goal-directed arm reaching movements to detect changes in fine hand motor skill after practice. Future research is needed to confirm the system's reliability and validity in assessing fine hand motor skills in patient populations.
... Currently, Kinect sensors are used in outdoor environments [29]- [32]. Kinect can generate 3D models of human skeletons in real time without any labels attached to the human body [33], and the measurement error of depth information is small [34], [35]. Therefore, the 3D data of the Kinect skeleton model can be utilized for gait recognition. ...
... 0-5 means the tracked body index, and -1 (0xFF) means no body is found. The processing speed is 30 frames per second [35]. Human joint extractions tracked by the Kinect sensor show an accuracy of less than 2 mm [47]. ...
Gait feature recognition refers to recognizing identities by collecting the characteristics of people when they walk. It shows the advantages of noncontact measurement, concealment, and nonimitability, and it also has good application value in monitoring, security, and company management. This paper utilizes Kinect to collect the three-dimensional coordinate data of human bones. Taking the spatial distances between the bone nodes as features, we solve the problem of placement and angle sensitivity of the camera. We design a fast and high-accuracy classifier based on the One-versus-one (OVO) and One-versus-rest (OVR) multiclassification algorithms derived from a support vector machine (SVM), which can realize the identification of persons without data records, and the number of classifiers is greatly reduced by design optimization. In terms of accuracy optimization, a filter based on n-fold Bernoulli theory is proposed to improve the classification accuracy of the multiclassifier. We select 20000 sets of data for fifty volunteers. Experimental results show that the design in this paper can effectively yield improved classification accuracy, which is 99.8%, and reduce the number of originally required classifiers by 91%-95%.
... In the following, we study the literature that attempted to assess the depth cameras as marker-less motion capture systems. (Schmitz et al., 2014) placed a jig (shown in Figure 1.11) in several static postures. A single Microsoft Kinect camera and a marker-based motion capture system record the object simultaneously. ...
... 10: The Kinetisense's setup for maker-less motion capture. Reproduced from Kinetisense. FIGURE 1.11: A jig with a ball-and-socket joint. This knee model was used to compare a Microsoft Kinect camera as a merker-less motion capture system with a marker-based motion capture system.(Schmitz et al., 2014) ...
L'analyse de la marche est la mesure et l'évaluation de la capacité de marche qui peut être utilisée pour l'évaluation des risques de chute ou comme outil de diagnostic et de pronostic pour des applications cliniques. Toutefois, malgré la valeur clinique, plusieurs difficultés attribuées à l'instrumentation de référence actuelle, les systèmes de capture du mouvement basés sur des marqueurs, limitent l'utilisation à grande échelle dans les applications cliniques. Les systèmes actuels sont coûteux et nécessitent un environnement de laboratoire contrôlé. La procédure de test est également longue. L'élimination des marqueurs réduirait considérablement le temps de préparation du patient et serait plus efficace. L'objectif de cette étude est de concevoir un système de capture de mouvement sans marqueur pour les applications cliniques. Les apprécents progrès réalisés dans le domaine de la vision par ordinateur et en particulier dans celui des réseaux neuronaux convolutifs, ont permis de poursuivre cet objectif. Le système conçu se compose de quatre caméras RGB et peut estimer la position des centres communs grâce à une approche d'apprentissage profond. À cette fin, un nouvel ensemble de données spécifiques a été collecté, incluant des sujets asymptomatiques et pathologiques. Pour évaluer la validité du système développé, ses performances sont évaluées par rapport à un système de capture de mouvement basé sur des marqueurs en termes d'erreurs de position des articulations et de paramètres de marche cliniquement pertinents. Les résultats démontrent le potentiel élevé du système conçu pour des applications cliniques.
... Vision sensor is one type of general sensors used in the rehabilitation assessment system. Kinect camera directly outputs joint angles as extracted features since it captures the depth of both filed and real-time image data (Schmitz, Ye, Shapiro, Yang, & Noehren, 2014). In a motion capture system, the joint angles and bodysegment orientation (Euler angles) are calculated using the body rotation coupled with marker trajectories. ...
... Angle frame is popular among the rule-based systems for real-time feedback (Clark et al., 2013;Bedregal et al., 2006;Hachaj & Ogiela, 2014) and some learning-based applications (Schmitz et al., 2014;Zhao et al., 2014) as it is a straightforward representation of human-readable movement. ...
Ageing causes loss of muscle strength, especially on the lower limbs, resulting in higher risk to injuries during functional activities. The path to recovery is through physiotherapy and adopt customized rehabilitation exercise to assist the patients. Hence, lowering the risk of incorrect exercise at home involves the use of biofeedback for physical rehabilitation patients and quantitative reports for clinical physiotherapy. This research topic has garnered much attention in recent years owing to the fast ageing population and the limited number of clinical experts. In this paper, the authors survey the existing works in exercise assessment and state identification. The evaluation results in the accuracy of 95.83% average classifying exercise motion state using the proposed raw signal. It confirmed that raw signals have more impact than using sensor-fused Euler and joint angles in the state identification prediction model.
... The use of 2D single-camera markerless motion capture, which consists of only one camera, involves the extraction of joint center positions from a picture or video via the application of 2D pose estimation algorithms [22]. Current research suggests an acceptable accuracy and reliability of single-camera markerless motion capture for joints with large ROM, such as knee and shoulder joints [24][25][26]. Most studies use the Microsoft Kinect video capture camera [27], as we did in our research. ...
Background: Total hip arthroplasty (THA) is one of the most cost-effective and successful procedures in orthopedics. However, assessing the post-operative range of motion (ROM) remains a challenge due to the limitations of traditional measurement methods. This study aimed to evaluate hip and spine ROM post-operatively and single-leg balance, using a single-camera markerless motion capture system, and compare outcomes with pre-operative ROM and with an age-matched healthy control group.
Methods: An interventional study was conducted from January 2018 to December 2021. Twenty patients with hip osteoarthritis underwent THA and were assessed using a single-camera markerless system (Kinetisense software). Measurements were taken one month pre-operatively and one year post-operatively.
Results: Significant improvements were observed in hip and lumbar spine ROM variables after THA. The most notable enhancements were in hip and spinal flexion. Compared to the control group, the THA group showed minor deficits in hip ROM, particularly in external rotation. Single-leg balance demonstrated improved stability post-operatively.
Conclusions: The single-camera markerless motion capture system offers a promising alternative for assessing hip and lumbar spine ROM, presenting potential advantages over manual goniometry and traditional 3D motion capture systems. Using this system for the evaluation of patients after THA, it seems that THA significantly enhances hip and lumbar spine ROM. Future research should focus on validating the accuracy of markerless systems.
... Three-dimensional marker trajectories were collected using a 10-camera motion capture system (MAC3D; Motion Analysis Corp., Santa Rosa, CA, United States) at 200 Hz. Trajectories were smoothed using a fourth-order Butterworth filter with an 8-Hz cut-off frequency (Chang et al., 2008;Schmitz et al., 2014). The center of the hip joint was estimated using the method described by Bell et al. (1987), Bell et al. (1990). ...
Knee sleeves are commonly used to address knee-related concerns, particularly in older individuals. Although previous studies have demonstrated their efficacy in improving gait and functional outcomes in knees with pathological conditions, the effectiveness of knee sleeves for improving gait characteristics in healthy older adults remains unclear. The harmonic ratio (HR), an index for assessing gait symmetry commonly used to discriminate between individuals with different functional levels, can be used to detect alterations in gait characteristics. This study investigated the effects of knee sleeves on gait symmetry in healthy older adults. Sixteen healthy community-dwelling older adults walked barefoot with and without knee sleeves at normal and fast speeds. Gait symmetry indices (HR and improved HR [iHR]) and spatiotemporal gait parameters were compared under different conditions. A significant interaction between knee condition and walking speed was observed for mean iHR in the anteroposterior direction (p = 0.006). A significant simple main effect of knee condition was found during fast walking, with a larger iHR with knee sleeves than without (p = 0.002). In the condition without knee sleeves, the iHR was significantly lower during fast walking than during normal walking (p = 0.035). Furthermore, a significant main effect of knee condition was observed for the variability of iHR in the anteroposterior direction, with a smaller variability when walking with knee sleeves than when walking without (p = 0.006). These results suggest that knee sleeves may enhance gait symmetry along the anteroposterior direction, particularly during fast walking, where symmetry disruption is more likely than walking at a comfortable pace. A significant reduction in gait symmetry variability also suggests a stabilizing effect on gait dynamics. These findings provide the first evidence supporting the efficacy of knee sleeves for improving gait symmetry. The use of knee sleeves could be a valuable option for restoring disrupted gait symmetry during fast walking, with potential implications for reducing the risk of falls.
... Gait can be detected for example through Dopler [12,13] or just sound [14]. Kinect depth-field cameras were also used, but usually from a side view [15][16][17][18]. Indeed all these approaches aim at maximizing identification, at odds with our approach. ...
This paper presents a preliminary study to assess the degree of anonymization provided by the use of depth field camera, for various degrees of pixelization. First the passage of 24 participants under a depth field camera was recorded. Each of the corresponding video was degraded with various levels of pixelization. Then the videos were shown to a subset of 6 participants, using a dedicated software which presents the videos in random order, starting with the lowest resolution. Each participant had to recognize themself, and in order to achieve this goal, could progressively improve the resolution. Our results question the fact that pixelization is the proper way to improve anonymity. Actually recognition seems to a large extend to be based on dynamic features rather than on the resolution of the picture. Besides we identify mostly 2 groups of responses: either the person can identify him/herself whatever the pixelization, or the recognition task is out of reach. Thus, the ability to use dynamic features could be person dependent. Further exploration would be useful to confirm this observation.
... However, the associated fitting process is time-and power-intensive; thus, such systems have restricted applicability. As a result, non-marker-based motion capture [35][36][37][38][39][40], which is based on high-accuracy RGB images and video processing techniques, has gained attention, with extensive research being conducted in the fields of deep learning and computer vision. ...
In this study, we proposed a novel transformer-based model with independent tokens for estimating three-dimensional (3D) human pose and shape from monocular videos, specifically focusing on its application in rehabilitation therapy. The main objective is to recover pixel-aligned rehabilitation-customized 3D human poses and body shapes directly from monocular images or videos, which is a challenging task owing to inherent ambiguity. Existing human pose estimation methods heavily rely on the initialized mean pose and shape as prior estimates and employ parameter regression with iterative error feedback. However, video-based approaches face difficulties capturing joint-level rotational motion and ensuring local temporal consistency despite enhancing single-frame features by modeling the overall changes in the image-level features. To address these limitations, we introduce two types of characterization tokens specifically designed for rehabilitation therapy: joint rotation and camera tokens. These tokens progressively interact with the image features through the transformer layers and encode prior knowledge of human 3D joint rotations (i.e., position information derived from large-scale data). By updating these tokens, we can estimate the SMPL parameters for a given image. Furthermore, we incorporate a temporal model that effectively captures the rotational temporal information of each joint, thereby reducing jitters in local parts. The performance of our method is comparable with those of the current best-performing models. In addition, we present the structural differences among the models to create a pose classification model for rehabilitation. We leveraged ResNet-50 and transformer architectures to achieve a remarkable PA-MPJPE of 49.0 mm for the 3DPW dataset.
... Several studies have indicated that MBased MoCap is considerably accurate [6][7][8][9][10]. Other studies [5,[11][12][13][14][15] have viewed that MLess MoCap is markedly appropriate. Among non-optical MoCap systems, inertial measurement unit (IMU) has been discussed as the best [16][17][18][19][20]. ...
Muscular skeletal disorder is a difficult challenge faced by the working population. Motion capture (MoCap) is used for recording the movement of people for clinical, ergonomic and rehabilitation solutions. However, knowledge barriers about these MoCap systems have made them difficult to use for many people. Despite this, no state-of-the-art literature review on MoCap systems for human clinical, rehabilitation and ergonomic analysis has been conducted. A medical diagnosis using AI applies machine learning algorithms and motion capture technologies to analyze patient data, enhancing diagnostic accuracy, enabling early disease detection and facilitating personalized treatment plans. It revolutionizes healthcare by harnessing the power of data-driven insights for improved patient outcomes and efficient clinical decision-making. The current review aimed to investigate: (i) the most used MoCap systems for clinical use, ergonomics and rehabilitation, (ii) their application and (iii) the target population. We used preferred reporting items for systematic reviews and meta-analysis guidelines for the review. Google Scholar, PubMed, Scopus and Web of Science were used to search for relevant published articles. The articles obtained were scrutinized by reading the abstracts and titles to determine their inclusion eligibility. Accordingly, articles with insufficient or irrelevant information were excluded from the screening. The search included studies published between 2013 and 2023 (including additional criteria). A total of 40 articles were eligible for review. The selected articles were further categorized in terms of the types of MoCap used, their application and the domain of the experiments. This review will serve as a guide for researchers and organizational management.
... Out of several technologies that capture body motion, marker-based motion capture has been the standard and most spatiotemporally accurate, though requiring the attachment of physical markers. Markerless video-based motion capture is a promising alternative used in situations where physical contact may not be possible [1][2][3]. It is good for the classification of movements but not yet with the spatial precision of marker systems. ...
Full-body motion capture is essential for the study of body movement. Video-based, markerless, mocap systems are, in some cases, replacing marker-based systems, but hybrid systems are less explored. We develop methods for coregistration between 2D video and 3D marker positions when precise spatial relationships are not known a priori. We illustrate these methods on three-ball cascade juggling in which it was not possible to use marker-based tracking of the balls, and no tracking of the hands was possible due to occlusion. Using recorded video and motion capture, we aimed to transform 2D ball coordinates into 3D body space as well as recover details of hand motion. We proposed four linear coregistration methods that differ in how they optimize ball-motion constraints during hold and flight phases, using an initial estimate of hand position based on arm and wrist markers. We found that minimizing the error between ball and hand estimate was globally suboptimal, distorting ball flight trajectories. The best-performing method used gravitational constraints to transform vertical coordinates and ball-hold constraints to transform lateral coordinates. This method enabled an accurate description of ball flight as well as a reconstruction of wrist movements. We discuss these findings in the broader context of video/motion capture coregistration.
... Eltoukhy (Eltoukhy et al., 2017) simultaneously used Kinect and Vicon system to record the results of Star Excursion Balance Test (SEBT), and found that the kinematic error of lower limbs was less than 5°except for the front plane angle of the posterior and lateral knee joints, which were 5-7°. According to Schmitz (Schmitz et al., 2014), the accuracy and precision of joint angle measurement in 3D skeleton model obtained from Kinect are equivalent to that of the mark-based system. Khoshelham (Khoshelham, 2012) believes that the point cloud data of Kinect sensor is able to provide acceptable accuracy by comparing it with the point cloud data of high-end laser scanner. ...
Introduction: Balance impairment is an important indicator to a variety of diseases. Early detection of balance impairment enables doctors to provide timely treatments to patients, thus reduce their fall risk and prevent related disease progression. Currently, balance abilities are usually assessed by balance scales, which depend heavily on the subjective judgement of assessors.
Methods: To address this issue, we specifically designed a method combining 3D skeleton data and deep convolutional neural network (DCNN) for automated balance abilities assessment during walking. A 3D skeleton dataset with three standardized balance ability levels were collected and used to establish the proposed method. To obtain better performance, different skeleton-node selections and different DCNN hyperparameters setting were compared. Leave-one-subject-out-cross-validation was used in training and validation of the networks.
Results and Discussion: Results showed that the proposed deep learning method was able to achieve 93.33% accuracy, 94.44% precision and 94.46% F1 score, which outperformed four other commonly used machine learning methods and CNN-based methods. We also found that data from body trunk and lower limbs are the most important while data from upper limbs may reduce model accuracy. To further validate the performance of the proposed method, we migrated and applied a state-of-the-art posture classification method to the walking balance ability assessment task. Results showed that the proposed DCNN model improved the accuracy of walking balance ability assessment. Layer-wise Relevance Propagation (LRP) was used to interpret the output of the proposed DCNN model. Our results suggest that DCNN classifier is a fast and accurate method for balance assessment during walking.
... 4,6,10,18 Recently, markerless motion capture systems have gained popularity due to efficiency, rapid data processing capabilities, and evidence-supporting accuracy and validity when compared with traditional marker-based systems. [23][24][25][26][27] Because of the rising popularity of markerless motion capture systems and the supported benefits of preparticipation movement assessments, the purpose of this study was to assess the bilateral shoulder ROM of firefighter trainees. Firefighters frequently suffer shoulder-related injuries, yet shoulder ROM data specific to firefighters remain scarce. ...
The unpredictable environments firefighters face paired with biomechanically compromising shoulder movements, such as overhead and lifting movements, place this population at an increased risk for shoulder injury. The purpose of this study was to assess firefighter trainees’ bilateral shoulder range of motion (ROM) using the Dynamic Athletic Research Institute Motion system. Retrospective anthropometric and ROM data for 31 male firefighter trainees were analyzed. Firefighter trainees’ mean shoulder ROM for bilateral external rotation, internal rotation, and extension were lower than previously published values. External rotation demonstrated the lowest percentage of trainees within normal ROM (left—6.67%, right—16.67%). Noting the susceptibility of upper extremity injuries among firefighters, establishing baseline ROM measurements for reference may improve musculoskeletal evaluations, training interventions, and injury rehabilitation.
... For instance, the hardware coefficient (such as, resolution, depth spacial precision, etc) and sensor accuracy (such as, average precision distribution and linearity, etc) were evaluated in [10], which clarifies the limitation of Kinect v2's depth measurement at the hardware level. The accuracy of gait tracking through single Kinect was evaluated in [11], which certifies the validation of a single Kinect to accurately measure lower extremity kinematics. Here, Kinect v2 was compared with a digital inclinometer with the measurement error of less than 0.5 • in sagittal and frontal plane rotation and less than 2 • in transverse plane rotation. ...
In this paper, a Kinect-based distributed and real-time motion capture system is developed. A trigonometric method is applied to calculate the relative position of Kinect v2 sensors with a calibration wand and register the sensors' positions automatically. By combining results from multiple sensors with a nonlinear least square method, the accuracy of the motion capture is optimized. Moreover, to exclude inaccurate results from sensors, a computational geometry is applied in the occlusion approach, which discovers occluded joint data. The synchronization approach is based on an NTP protocol that synchronizes the time between the clocks of a server and clients dynamically, ensuring that the proposed system is a real-time system. Experiments for validating the proposed system are conducted from the perspective of calibration, occlusion, accuracy, and efficiency. Furthermore, to demonstrate the practical performance of our system, a comparison of previously developed motion capture systems (the linear trilateration approach and the geometric trilateration approach) with the benchmark OptiTrack system is conducted, therein showing that the accuracy of our proposed system is and 24.1% better than the two aforementioned trilateration systems, respectively.
... The nearer to the camera the subject is, the higher the accuracy is. Instead, Ref. [31] proposed a calculation pipeline based on the data recorded by a single camera, and they compared it with the positions contained in a database. ...
Nowadays, the current market trend is oriented toward increasing mass customization, meaning that modern production systems have to be able to be flexible but also highly productive. This is due to the fact that we are still living in the so-called Industry 4.0, with its cornerstone of high-productivity systems. However, there is also a migration toward Industry 5.0 that includes the human-centered design of the workplace as one of its principles. This means that the operators have to be put in the center of the design techniques in order to maximize their wellness. Among the wide set of new technologies, collaborative robots (cobots) represent one such technology that modern production systems are trying to integrate, because of their characteristic of working directly with the human operators, allowing for a mix of the flexibility of the manual systems with the productivity of the automated ones. This paper focuses on the impact that these technologies have on different levels within a production plant and on the improvement of the collaborative experience. At the workstation level, the control methodologies are investigated and developed: technologies such as computer vision and augmented reality can be applied to aid and guide the activities of the cobot, in order to obtain the following results. The first is an increase in the overall productivity generated by the reduction of idle times and safety stops and the minimization of the effort required to the operator during the work. This can be achieved through a multiobjective task allocation which aims to simultaneoulsy minimize the makespan, for productivity requirements, and the operator’s energy expenditure and mental workload, for wellness requirements. The second is a safe, human-centered, workspace in which collisions can be avoided in real time. This can be achieved by using real-time multicamera systems and skeleton tracking to constantly know where the operator is in the work cell. The system will offer the possibility of directing feedback based on the discrepancies between the physical world and the virtual models in order to dynamically reallocate the tasks to the resources if the requirements are not satisfied anymore. This allows the application of the technology to sectors that require constant process control, improving also the human–robot interaction: the human operator and the cobot are not merely two single resources working in the same cell, but they can achieve a real human–robot collaboration. In this paper, a framework is preented that allows us to reach the different aforementioned goals.
... Typically, joint angles are obtained using a conventional marker-based motion tracking system (MTS); however, the need to use an array of cameras makes their implementation unrealistic in the field [13]. Correspondingly, a low-cost and markerless Kinect sensor can be an affordable 3D sensing device and provide an accessible alternative to MTSs for working posture assessment in field studies for companies that do not have the financial capacity to afford high-priced systems and for full-time ergonomists to use the assessment tools properly. ...
A trunk-twisting posture is strongly associated with physical discomfort. Measurement of joint kinematics to assess physical exposure to injuries is important. However, using a single Kinect sensor to track the upper-limb joint angle trajectories during twisting tasks in the workplace is challenging due to sensor view occlusions. This study provides and validates a simple method to optimally select the upper-limb joint angle data from two Kinect sensors at different viewing angles during the twisting task, so the errors of trajectory estimation can be improved. Twelve healthy participants performed a rightward twisting task. The tracking errors of the upper-limb joint angle trajectories of two Kinect sensors during the twisting task were estimated based on concurrent data collected using a conventional motion tracking system. The error values were applied to generate the error trendlines of two Kinect sensors using third-order polynomial regressions. The intersections between two error trendlines were used to define the optimal data selection points for data integration. The finding indicates that integrating the outputs from two Kinect sensor datasets using the proposed method can be more robust than using a single sensor for upper-limb joint angle trajectory estimations during the twisting task.
... First of all, we proposed cut-off values to distinguish PD from HCs and PD with postural abnormalities from those without via global assessment of postures. Moreover, the IPA was obtained based on objective kinematic features derived from Kinect depth camera and computer algorithm which was accurate and repeatable 30,43 . In addition, ...
Postural abnormalities are common disabling motor complications affecting patients with Parkinson’s disease (PD). We proposed a summary index for postural abnormalities (IPA) based on Kinect depth camera and explored the clinical value of this indicator. Seventy individuals with PD and thirty age-matched healthy controls (HCs) were enrolled. All participants were tested using a Kinect-based system with IPA automatically obtained by algorithms. Significant correlations were detected between IPA and the Movement Disorder Society-Sponsored Revision of the Unified Parkinson’s Disease Rating Scale (MDS-UPDRS) total score (rs = 0.369, p = 0.002), MDS-UPDRS-III total score (rs = 0.431, p < 0.001), MDS-UPDRS-III 3.13 score (rs = 0.573, p < 0.001), MDS-UPDRS-III-bradykinesia score (rs = 0.311, p = 0.010), the 39-item Parkinson’s Disease Questionnaire (PDQ-39) (rs = 0.272, p = 0.0027) and the Berg Balance Scale (BBS) score (rs = −0.350, p = 0.006). The optimal cut-off value of IPA for distinguishing PD from HCs was 12.96 with a sensitivity of 97.14%, specificity of 100.00%, area under the curve (AUC) of 0.999 (0.997–1.002, p < 0.001), and adjusted AUC of 0.998 (0.993–1.000, p < 0.001). The optimal cut-off value of IPA for distinguishing between PD with and without postural abnormalities was 20.14 with a sensitivity, specificity, AUC and adjusted AUC of 77.78%, 73.53%, 0.817 (0.720–0.914, p < 0.001), and 0.783 (0.631–0.900, p < 0.001), respectively. IPA was significantly correlated to the clinical manifestations of PD patients, and could reflect the global severity of postural abnormalities in PD with important value in distinguishing PD from HCs and distinguishing PD with postural abnormalities from those without.
... The Microsoft Kinect sensor is well suited for indoor and outdoor environments because of the markerless motion analysis, easy accessibility of sensor data, and costeffectiveness. The Kinect sensor can generate 3D skeleton data at the speed of 30 frames per second [30]. Moreover, the extraction of the body joints tracked by the Kinect sensor shows the accuracy and precision of less than 2 mm [31]. ...
Over the past decade, gait recognition had gained a lot of attention in various research and industrial domains. These include remote surveillance, border control, medical rehabilitation, emotion detection from posture, fall detection, and sports training. The main advantages of identifying a person by their gait include unobtrusiveness, acceptance, and low costs. This paper proposes a convolutional neural network KinectGaitNet for Kinect-based gait recognition. The 3D coordinates of each of the body joints over the gait cycle are transformed to create a unique input representation. The proposed KinectGaitNet is trained directly using the 3D input representation without the necessity of the handcrafted features. The KinectGaitNet design allows avoiding gait cycle resampling, and the residual learning method ensures high accuracy without the degradation problem. The proposed deep learning architecture surpasses the recognition performance of all state-of-the-art methods for Kinect-based gait recognition by achieving 96.91% accuracy on UPCV and 99.33% accuracy on the KGB dataset. The method is the first, to the best of our knowledge, deep learning-based architecture that is based on a unique 3D input representation of joint coordinates. It achieves performance higher than previous traditional and deep learning methods, with fewer parameters and shorter inference time.
... They introduced the "stiffness-strength-stability coupling model" between the hydraulic support and surrounding rock, which provides an approach to dynamically analyze and predict hydraulic support loads in longwall mining faces. Some scholars have studied the evolution mechanism of coal and rock dynamics in coal mining processes (Schmitz et al. 2014;Dennis 2019;Qi et al. 2020;Mu et al. 2020). These works revealed the mechanism of dynamic disaster occurrence, proposed prevention and control concepts based on source separation and classification, and developed supporting equipment to provide theoretical, technical, and equipment support for dynamic disaster prevention and control. ...
Hydraulic support is the primary equipment used for surrounding rock control at fully mechanized mining faces. The load, location, and attitude of the hydraulic support are important sets of basis data to predict roof disasters. This paper summarized and analyzed the status of coal mine safety accidents and the primary influencing factors of roof disasters. This work also proposed monitoring characteristic parameters of roof disasters based on support posture-load changes, such as the support location and support posture. The data feature decomposition method of the additive model was used with the monitoring load data of the hydraulic support in the Yanghuopan coal mine to effectively extract the trend, cycle period, and residuals, which provided the period weighting characteristics of the longwall face. The autoregressive, long-short term memory, and support vector regression algorithms were used to model and analyze the monitoring data to realize single-point predictions. The seasonal autoregressive integrated moving average (SARIMA) and autoregressive integrated moving average (ARIMA) models were adopted to predict the support cycle load of the hydraulic support. The SARIMA model is shown to be better than the ARIMA model for load predictions in one support cycle, but the prediction effect of these two algorithms over a fracture cycle is poor. Therefore, we proposed a hydraulic support load prediction method based on multiple data cutting and a hydraulic support load template library. The constructed technical framework of the roof disaster intelligent prediction platform is based on this method to perform predictions and early warnings of roof disasters based on the load and posture monitoring information from the hydraulic support.
... These methods enable the automated tracking of kinematic variables using simple digital cameras, allowing greater portability and the collection of data in more ecological settings. A growing number of commercial markerless motion capture systems have become available, with several showing promising accuracy/reliability. [2][3][4][5][6] However, most of these systems require expensive subscriptions and are limited, in a sense, to capturing 3D kinematics, in that 3D kinematics will always require multicamera systems, calibrated to a capture volume. ...
Several open-source platforms for markerless motion capture offer the ability to track 2-dimensional (2D) kinematics using simple digital video cameras. We sought to establish the performance of one of these platforms, DeepLabCut. Eighty-four runners who had sagittal plane videos recorded of their left lower leg were included in the study. Data from 50 participants were used to train a deep neural network for 2D pose estimation of the foot and tibia segments. The trained model was used to process novel videos from 34 participants for continuous 2D coordinate data. Overall network accuracy was assessed using the train/test errors. Foot and tibia angles were calculated for 7 strides using manual digitization and markerless methods. Agreement was assessed with mean absolute differences and intraclass correlation coefficients. Bland–Altman plots and paired t tests were used to assess systematic bias. The train/test errors for the trained network were 2.87/7.79 pixels, respectively (0.5/1.2 cm). Compared to manual digitization, the markerless method was found to systematically overestimate foot angles and underestimate tibial angles ( P < .01, d = 0.06–0.26). However, excellent agreement was found between the segment calculation methods, with mean differences ≤1° and intraclass correlation coefficients ≥.90. Overall, these results demonstrate that open-source, markerless methods are a promising new tool for analyzing human motion.
... Although Kinect 1 is a well-established system, when compared to gold standard techniques, it provides only basic motion capture capabilities such as collecting temporal gait parameters, estimating single joint angles, or assessing postural control during reaching and balance tasks (Clark et al., 2012;Schmitz et al., 2014). In kinematic gait recordings, the system generally underestimates joint flexion and overestimates extension during walking in the sagittal plane. ...
The understanding of locomotion in neurological disorders requires technologies for quantitative gait analysis. Numerous modalities are available today to objectively capture spatiotemporal gait and postural control features. Nevertheless, many obstacles prevent the application of these technologies to their full potential in neurological research and especially clinical practice. These include the required expert knowledge, time for data collection, and missing standards for data analysis and reporting. Here, we provide a technological review of wearable and vision-based portable motion analysis tools that emerged in the last decade with recent applications in neurological disorders such as Parkinson's disease and Multiple Sclerosis. The goal is to enable the reader to understand the available technologies with their individual strengths and limitations in order to make an informed decision for own investigations and clinical applications. We foresee that ongoing developments toward user-friendly automated devices will allow for closed-loop applications, long-term monitoring, and telemedical consulting in real-life environments.
... However, the accuracy assessment of Kinect as a motion capture system with a ground-truth measure and how it compares to commercial motion capture systems to measure lower-extremity kinematics remains understudied. A digital inclinometer was employed to measure the angles of a ball-and-socket joint as ground-truth by [39]. The accuracy and repeatability of Kinect v1 in capturing static postures of the joint were qualified, and the performance was compared with a markerbased motion capture system. ...
... Conversely, a single camera marker-less motion capture system in [14] had developed for estimating the joint angles. A marker-less system based on the OpenPose library was introduced for gait analysis [15]. ...
Elderly health monitoring, rehabilitation training, and sport supervision could benefit from continuous assessment of joint angle, and angular velocity to identify the joint movement patterns. However, most of the measurement systems are designed based on special kinematic sensors to estimate angular velocities. The study aims to measure the lower limb joint angular velocity based on a 2D vision camera system during squat and walking on treadmill action using deep convolution neural network (CNN) architecture. Experiments were conducted on 12 healthy adults, and six digital cameras were used to capture the videos of the participant actions in lateral and frontal view. The normalized cross-correlation (Ccnorm) analysis was performed to obtain a degree of symmetry of the ground truth and estimated angular velocity waveform patterns. Mean Ccnorm for angular velocity estimation by deep CNN model has higher than 0.90 in walking on the treadmill and 0.89 in squat action. Furthermore, joint-wise angular velocities at the hip, knee, and ankle joints were observed and compared. The proposed system gets higher estimation performance under the lateral view and the frontal view of the camera. This study potentially eliminates the requirement of wearable sensors and proves the applicability of using video-based system to measure joint angular velocities during squat and walking on a treadmill actions.
... Therefore, this study is a feasibility study towards validating the SWING test as a routine clinical hip RoM assessment. Subsequent work could aim to provide low cost, accessible and easy-to-use assessment tools for performing the SWING test in a clinical context, such as the inertial measurement units [26] or single camera markerless motion capture system [27]. Future studies will also evaluate the SWING test's reliability on pathological populations such as LBP patients. ...
Background
Clinical assessment of sagittal plane hip mobility is usually performed using the Modified Thomas Test (for extension) and the Straight-Leg-Raise (for flexion) with a goniometer. These tests have limited reliability, however. An active swinging leg movement test (the SWING test), assessed using 3D motion analysis, could provide an alternative to these passive clinical tests.
Research question
Is the SWING test a more reliable alternative to evaluate hip mobility, in comparison to the clinical extension and flexion tests?
Methods
Ten asymptomatic adult participants were evaluated by two investigators over three sessions. Participants performed 10 maximal hip extensions and flexions, with both legs straight and no trunk movement (the SWING test). Hip kinematics was assessed using a 3D motion analysis system. Maximal and minimal hip angles were calculated for each swing and represented maximal hip flexion (SWING flexion) and extension (SWING extension), respectively. The Modified Thomas Test and Straight-Leg-Raise were repeated 3 times for each leg. On the first day, both investigators performed all the tests (SWING + Modified Thomas Test + Straight-Leg-Raise). A week later, a single investigator repeated all the tests. Inter-rater, intra-rater, within-day and between-day reliability were evaluated using intra-class correlation.
Results
Intra-class correlation coefficients for all the tests were superior to 0.8, except for the Modified Thomas Test’s intra-rater, between-day (intra-class correlation 0.673) and the Straight-Leg-Raise’s inter-rater, within-day (intra-class correlation 0.294). The SWING test always showed a higher intra-class correlation coefficient than the passive clinical tests. The only significant correlation found was for the Straight-Leg-Raise and SWING flexion (r = 0.48; P < 0.001).
Significance
The SWING test seems to be an alternative to existing passive clinical tests, offering better reliability for assessing sagittal plane hip mobility.
... A three-dimensional motion-tracking system (Myo-Motion; Noraxon Inc., Scottsdale, AZ, USA) sampling at 200 Hz was used to analyze the kinematic variables. The accuracy and reliability of MyoMotion system were reported by Schmitz et al. [34]. Myo-Motion Research inertial sensors were placed according to the rigid-body model, with 9 segments defined in the MR3 software as shown in Fig. 2. Sensors were placed to the surface of the 7th cervical spinous process (upper thoracic), the 12th thoracic spinous process (lower thoracic), at the midpoint of the posterior superior iliac spine, the frontal attachment on lower quadrant of quadriceps, slightly above the patella, shanks (frontal on the tibia), and top of the upper foot, slightly below the ankle. ...
Background
This study sought to determine the effects of a 6-week neuromuscular training (NMT) and NMT plus external focus (NMT plus EF) programs on trunk and lower extremity inter-segmental movement coordination in active individuals at risk of injury.
Methods
Forty-six active male athletes (controls = 15, NMT = 16, NMT plus EF = 15) participated (age = 23.26 ± 2.31 years) in this controlled, laboratory study. Three-dimensional kinematics were collected during a drop vertical jump (DVJ). A continuous relative phase (CRP) analysis quantified inter-segmental coordination of the: (1) thigh (flexion/extension)—shank (flexion/extension), (2) thigh (abduction/adduction)—shank (flexion/extension), (3) thigh (abduction/adduction)—trunk (flexion/extension), and (4) trunk (flexion/extension)—pelvis (posterior tilt/anterior tilt). Analysis of covariance compared biomechanical data between groups.
Results
After 6 weeks, inter-segmental coordination patterns were significantly different between the NMT and NMT plus EF groups (p < 0.05). No significant differences were observed in CRP for trunk-pelvis coupling comparing between NMT and NMT plus EF groups (p = 0.134), while significant differences were observed CRP angle of the thigh-shank, thigh-trunk couplings (p < 0.05).
Conclusions
Trunk and lower extremity movement coordination were more in-phase during DVJ in the NMT plus EF compared to NMT in active individuals at risk of anterior cruciate ligament injury.
... They are characterized by an accuracy of less than one millimeter error in the detection of markers [31,32]. In addition to their use in motion analysis, 3DMAS are also used as a gold standard in verifying the accuracy of new methods of motion detection [23,24,31,[33][34][35]. ...
Demographic changes associated with an expanding and aging population will lead to an increasing number of orthopedic surgeries, such as joint replacements. To support patients’ home exercise programs after total hip replacement and completing subsequent inpatient rehabilitation, a low-cost, smartphone-based augmented reality training game (TG) was developed. To evaluate its motion detection accuracy, data from 30 healthy participants were recorded while using the TG. A 3D motion analysis system served as reference. The TG showed differences of 18.03 mm to 24.98 mm along the anatomical axes. Surveying the main movement direction of the implemented exercises (squats, step-ups, side-steps), differences between 10.13 mm to 24.59 mm were measured. In summary, the accuracy of the TG’s motion detection is sufficient for use in exergames and to quantify progress in patients’ performance. Considering the findings of this study, the presented exer-game approach has potential as a low-cost, easily accessible support for patients in their home exercise program.
... These methods enable the automated tracking of kinematic variables using simple digital cameras, allowing greater portability and the collection of data in more ecological settings. A growing number of commercial markerless motion capture systems have become available, with several showing promising accuracy/reliability. [2][3][4][5][6] However, most of these systems require expensive subscriptions and are limited, in a sense, to capturing 3D kinematics, in that 3D kinematics will always require multicamera systems, calibrated to a capture volume. ...
... A fully instrumented participant can be seen in Figure 3a. It is assumed that the resulting joint angles from Qualisys are more accurate, and they were therefore be used as the ground truth angles [23,24]. ...
The reproduction and simulation of workplaces, and the analysis of body postures during work processes, are parts of ergonomic risk assessments. A commercial virtual reality (VR) system offers the possibility to model complex work scenarios as virtual mock-ups and to evaluate their ergonomic designs by analyzing motion behavior while performing work processes. In this study a VR tracking sensor system (HTC Vive tracker) combined with an inverse kinematic model (Final IK) was compared with a marker-based optical motion capture system (Qualisys). Marker-based optical motion capture systems are considered the gold standard for motion analysis. Therefore, Qualisys was used as the ground truth in this study. The research question to be answered was how accurately the HTC Vive System combined with Final IK can measure joint angles used for ergonomic evaluation. Twenty-six subjects were observed simultaneously with both tracking systems while performing 20 defined movements. Sixteen joint angles were analyzed. Joint angle deviations between ±6∘ and ±42∘ were identified. These high deviations must be considered in ergonomic risk assessments when using a VR system. The results show that commercial low-budget tracking systems have the potential to map joint angles. Nevertheless, substantial weaknesses and inaccuracies in some body regions must be taken into account. Recommendations are provided to improve tracking accuracy and avoid systematic errors.
... https://doi.org/10.1371/journal.pcbi.1008935.g008 [20,[33][34][35][36][37][38], and a variety of other technologies [18,[39][40][41][42]. We did not directly compare the results of our OpenPose analyses to results of any of these other markerless approaches, and thus we are hesitant to speculate about the relative accuracy of our approach against others. ...
Human gait analysis is often conducted in clinical and basic research, but many common approaches (e.g., three-dimensional motion capture, wearables) are expensive, immobile, data-limited, and require expertise. Recent advances in video-based pose estimation suggest potential for gait analysis using two-dimensional video collected from readily accessible devices (e.g., smartphones). To date, several studies have extracted features of human gait using markerless pose estimation. However, we currently lack evaluation of video-based approaches using a dataset of human gait for a wide range of gait parameters on a stride-by-stride basis and a workflow for performing gait analysis from video. Here, we compared spatiotemporal and sagittal kinematic gait parameters measured with OpenPose (open-source video-based human pose estimation) against simultaneously recorded three-dimensional motion capture from overground walking of healthy adults. When assessing all individual steps in the walking bouts, we observed mean absolute errors between motion capture and OpenPose of 0.02 s for temporal gait parameters (i.e., step time, stance time, swing time and double support time) and 0.049 m for step lengths. Accuracy improved when spatiotemporal gait parameters were calculated as individual participant mean values: mean absolute error was 0.01 s for temporal gait parameters and 0.018 m for step lengths. The greatest difference in gait speed between motion capture and OpenPose was less than 0.10 m s⁻¹. Mean absolute error of sagittal plane hip, knee and ankle angles between motion capture and OpenPose were 4.0°, 5.6° and 7.4°. Our analysis workflow is freely available, involves minimal user input, and does not require prior gait analysis expertise. Finally, we offer suggestions and considerations for future applications of pose estimation for human gait analysis.
... Finally, instead of tracking a particular object or body markers, for which marker occlusion could be an issue, one approach consists of using markerless 3D motion capture systems (Mündermann, Corazza, & Andriacchi, 2006), using deep learning-based approaches with 2D or 3D pose estimation (Iskakov, Burkov, Lempitsky, & Malkov, 2019;Pavlakos, Zhou, & Daniilidis, 2018;Pavllo, Feichtenhofer, Grangier, & Auli, 2019) or RGB-depth cameras such as Microsoft Kinect™ (Clark et al., 2012;Gao, Yu, Zhou, & Du, 2015;Liddy et al., 2017;Pfister, West, Bronner, & Noah, 2014;Schmitz, Ye, Shapiro, Yang, & Noehren, 2014). Liddy et al. (2017) investigated the temporal shift between the Microsoft Kinect TM system and the optokinetic Vicon system during a bi-manual coordination task at frequencies ranging between 1 and 3.33 Hz. ...
Introduction
Advances in motion capture technology include markerless systems to facilitate valid data collection. Recently, the technological reliability of this technology has been reported for human movement assessments. To further understand sources of potential error, biological reliability must also be determined. The aim of this study was to determine the day-to-day reliability for a three-dimensional markerless motion capture (MMC) system to quantify 4 movement analysis composite scores, and 81 kinematic variables.
Methods
Twenty-two healthy men ( n = 11; X ¯ ± SD ; age = 23.0 ± 2.6 years, height = 180.4.8 cm, weight = 80.4 ± 7.3 kg) and women ( n = 11; age = 20.8 ± 1.1 years, height = 172.2 ± 7.4 cm, weight = 68.0 ± 7.3 kg) participated in this study. All subjects performed 4 standardized test batteries consisting of 14 different movements on four separate days. A three-dimensional MMC system (DARI Motion, Lenexa, KS) using 8 cameras surrounding the testing area was used to quantify movement characteristics. 1 × 4 RMANOVAs were used to determine significant differences across days for the composite movement analysis scores, and RM-MANOVAs were used to determine test day differences for the kinematic data ( p < 0.05). Intraclass correlation coefficients (ICCs) were reported for all variables to determine test reliability. To determine biological variability, mean absolute differences from previously reported technological variability data were subtracted from the total variability data from the present study.
Results
No differences were observed for any composite score (i.e., athleticism, explosiveness, quality, readiness; or any of the 81 kinematic variables. Furthermore, 84 of 85 measured variables exhibited good to excellent ICCs (0.61–0.99). When compared to previously reported technological variability data, 62.3% of item variability was due to biological variability, with 66 of 85 variables exhibiting biological variability as the primary source of error (i.e., >50% total variability).
Discussion
Combined, these findings effectively add to the body of literature suggesting sufficient reliability for MMC solutions in capturing kinematic features of human movement.
Markerless motion capture system and X-ray fluoroscopy as two markerless measurement systems were introduced to the application method in sports biomechanical areas. An overview of the technological process, data accuracy, suggested movements, and recommended body parts were explained. The markerless motion capture system consists of four parts: camera, body model, image feature, and algorithms. Even though the markerless motion capture system seems promising, it is not yet known whether these systems can be used to achieve the required accuracy and whether they can be appropriately used in sports biomechanics and clinical research. The biplane fluoroscopy technique analyzes motion data by collecting, image calibrating, and processing, which is effective for determining small joint kinematic changes and calculating joint angles. The method was used to measure walking and jumping movements primarily because of the experimental conditions and mainly to detect the data of lower limb joints.
Kinect merupakan perangkat sensor motion capture yang mulanya dipakai untuk video game controller, tapi kini digunakan untuk berbagai macam keperluan yang membutuhkan sensor gerak. Kinect relatif murah dan mudah untuk didapatkan dan dipakai sebagai motion capture untuk menggerakkan karakter digital 3D. Dalam penelitian ini, uji coba dilakukan terhadap kemampuan Kinect dalam set-up sistem motion capture yang sesederhana mungkin dan mudah untuk dapat diakses oleh awam. Dari hasil uji coba didapatkan hasil dan saran atau anjuran pemakaian dan set-up yang paling efektif untuk dapat menggerakkan sebuah karakter digital 3D. Diharapkan hasil ini dapat mendorong penggunaan potensi Kinect lebih maksimal.
Objective
Advances in computer vision make it possible to combine low-cost cameras with algorithms, enabling biomechanical measures of body function and rehabilitation programs to be performed anywhere. We evaluated a computer vision system's accuracy and concurrent validity for estimating clinically relevant biomechanical measures.
Design
Cross-sectional study.
Setting
Laboratory.
Participants
Thirty-one healthy participants and 31 patients with axial spondyloarthropathy.
Intervention
A series of clinical functional tests (including the gold standard Bath Ankylosing Spondylitis Metrology Index tests). Each test was performed twice: the first performance was recorded with a camera, and a computer vision algorithm was used to estimate variables. During the second performance, a clinician measured the same variables manually.
Main measures
Joint angles and inter-limb distances. Clinician measures were compared with computer vision estimates.
Results
For all tests, clinician and computer vision estimates were correlated ( r ² values: 0.360–0.768). There were no significant mean differences between methods for shoulder flexion (left: 2 ± 14° (mean ± standard deviation), t = 0.99, p < 0.33; right: 3 ± 15°, t = 1.57, p < 0.12), side flexion (left: − 0.5 ± 3.1 cm, t = −1.34, p = 0.19; right: 0.5 ± 3.4 cm, t = 1.05, p = 0.30) and lumbar flexion ( − 1.1 ± 8.2 cm, t = −1.05, p = 0.30). For all other movements, significant differences were observed, but could be corrected using a systematic offset.
Conclusion
We present a computer vision approach that estimates distances and angles from clinical movements recorded with a phone or webcam. In the future, this approach could be used to monitor functional capacity and support physical therapy management remotely.
The joint angular velocity during daily life exercises is an important clinical outcome for injury risk index, rehabilitation progress monitoring and athlete's performance evaluation. Recently, wearable sensors have been widely used to monitor lower limb kinematics. However, these sensors are difficult and inconvenient to use in daily life. To mitigate these limitations, this study proposes a vision-based system for estimating lower limb joint kinematics using a deep convolution neural network with bi-directional long-short term memory and gated recurrent unit network. The normalized correlation coefficient, and the mean absolute error were computed between the ground truth obtained from the optical motion capture system and estimated joint angular velocities using proposed models. The estimated results show a highest correlation 0.93 in squat and 0.92 in walking on treadmill action. Furthermore, independent model for each joint angular velocity at the hip, knee, and ankle were analyzed and compared. Among the three joint angular velocities, knee joint has a best estimated accuracy (0.96 in squat and 0.96 in walking on the treadmill). The proposed models show higher estimation accuracy under both the lateral and the frontal view regardless of the camera positions and angles. This study proves the applicability of using sensor free vision-based system to monitor the lower limb kinematics during home workouts for healthcare and rehabilitation.
Vision-based human joint angle estimation is essential for remote and continuous health monitoring. Most vision-based angle estimation methods use the locations of human joints extracted using optical motion cameras, depth cameras, or human pose estimation models. This study aimed to propose a reliable and straightforward approach with deep learning networks for knee and elbow flexion/extension angle estimation from the RGB video. Fifteen healthy participants performed four daily activities in this study. The experiments were conducted with four different deep learning networks, and the networks took nine subsequent frames as input while output was knee and elbow joint angles extracted from an optical motion capture system for each frame. The BiLSTM network-based joint angles estimator can estimate both joint angles with a correlation of 0.955 for knee and 0.917 for elbow joints regardless of the camera view angles.
The hand region is reported as the most common injury site in boxing, with more observed time loss than any other area in this sport. The amount of wrist motion, specifically flexion, has been described as contributing to these injuries, yet no literature is available to quantify wrist kinematics in boxing. This is the first paper describing wrist motion on impact in boxing. Utilising an electromagnetic tracking system, two types of shots were assessed, Jab (straight arm) and Hook (bent arm), during in-vivo testing procedures with 29 elite boxers. For both shots, flexion and ulnar deviation occurred concurrent on impact, with an M and SD of 9.3 ± 1.9° and 4.7 ± 1.2° respectively for Jab shots, and 5.5 ± 1.1° and 3.3 ± 1.1° respectively for Hook shots, supporting dart throwing motion at the wrist. For both Jab & Hook, wrist motion on impact occurred within >30% and >20% respectively of total available active range of motion, with wrist angles greater in both flexion (t = 9.0, p < 0.001, d = 1.7) and ulnar deviation (t = 8.4, p < 0.001, d = 1.6) for Jab compared to Hook shots. The study provides novel and quantifiable information regarding wrist kinematics during the impact phase of punching and potentially an improved understanding of injury mechanisms in boxing.
Background: Several factors, including the aging population and the recent corona pandemic, have increased the need for cost effective,
easy-to-use and reliable telerehabilitation services. Computer vision-based marker-less human pose estimation is a promising variant of telerehabilitation and is currently an intensive research topic. It has attracted significant interest for detailed motion analysis, as it does not need arrangement of external fiducials while capturing motion data from images. This is promising for rehabilitation applications, as they enable analysis and
supervision of clients’ exercises and reduce clients’ need for visiting physiotherapists in person. However, development of a marker-less motion
analysis system with precise accuracy for joint identification, joint angle measurements and advanced motion analysis is an open challenge.
Objectives: The main objective of this paper is to provide a critical overview of recent computer vision-based marker-less human pose
estimation systems and their applicability for rehabilitation application. An overview of some existing marker-less rehabilitation applications
is also provided.
Methods: This paper presents a critical review of recent computer vision-based marker-less human pose estimation systems with focus
on their provided joint localization accuracy in comparison to physiotherapy requirements and ease of use. The accuracy, in terms of the
capability to measure the knee angle, is analysed using simulation.
Results: Current pose estimation systems use 2D, 3D, multiple and single view-based techniques. The most promising techniques from
a physiotherapy point of view are 3D marker-less pose estimation based on a single view as these can perform advanced motion analysis
of the human body while only requiring a single camera and a computing device. Preliminary simulations reveal that some proposed systems
already provide a sufficient accuracy for 2D joint angle estimations.
Conclusions: Even though test results of different applications for some proposed techniques are promising, more rigour testing is
required for validating their accuracy before they can be widely adopted in advanced rehabilitation applications.
Keywords: Computer vision, marker-less, motion analysis, telerehabilitation
To evaluate the postures in ergonomics applications, studies have proposed the use of low-cost, marker-less, and portable depth camera-based motion tracking systems (DCMTSs) as a potential alternative to conventional marker-based motion tracking systems (MMTSs). However, a simple but systematic method for examining the estimation errors of various DCMTSs is lacking. This paper proposes a benchmarking method for assessing the estimation accuracy of depth cameras for full-body landmark location estimation. A novel alignment board was fabricated to align the coordinate systems of the DCMTSs and MMTSs. The data from an MMTS were used as a reference to quantify the error of using a DCMTS to identify target locations in a 3-D space. To demonstrate the proposed method, the full-body landmark location tracking errors were evaluated for a static upright posture using two different DCMTSs. For each landmark, we compared each DCMTS (Kinect system and RealSense system) with an MMTS by calculating the Euclidean distances between symmetrical landmarks. The evaluation trials were performed twice. The agreement between the tracking errors of the two evaluation trials was assessed using intraclass correlation coefficient (ICC). The results indicate that the proposed method can effectively assess the tracking performance of DCMTSs. The average errors (standard deviation) for the Kinect system and RealSense system were 2.80 (1.03) cm and 5.14 (1.49) cm, respectively. The highest average error values were observed in the depth orientation for both DCMTSs. The proposed method achieved high reliability with ICCs of 0.97 and 0.92 for the Kinect system and RealSense system, respectively.
Background:
Several factors, including the aging population and the recent corona pandemic, have increased the need for cost effective, easy-to-use and reliable telerehabilitation services. Computer vision-based marker-less human pose estimation is a promising variant of telerehabilitation and is currently an intensive research topic. It has attracted significant interest for detailed motion analysis, as it does not need arrangement of external fiducials while capturing motion data from images. This is promising for rehabilitation applications, as they enable analysis and supervision of clients' exercises and reduce clients' need for visiting physiotherapists in person. However, development of a marker-less motion analysis system with precise accuracy for joint identification, joint angle measurements and advanced motion analysis is an open challenge.
Objectives:
The main objective of this paper is to provide a critical overview of recent computer vision-based marker-less human pose estimation systems and their applicability for rehabilitation application. An overview of some existing marker-less rehabilitation applications is also provided.
Methods:
This paper presents a critical review of recent computer vision-based marker-less human pose estimation systems with focus on their provided joint localization accuracy in comparison to physiotherapy requirements and ease of use. The accuracy, in terms of the capability to measure the knee angle, is analysed using simulation.
Results:
Current pose estimation systems use 2D, 3D, multiple and single view-based techniques. The most promising techniques from a physiotherapy point of view are 3D marker-less pose estimation based on a single view as these can perform advanced motion analysis of the human body while only requiring a single camera and a computing device. Preliminary simulations reveal that some proposed systems already provide a sufficient accuracy for 2D joint angle estimations.
Conclusions:
Even though test results of different applications for some proposed techniques are promising, more rigour testing is required for validating their accuracy before they can be widely adopted in advanced rehabilitation applications.
To understand the dynamic patterns of behaviours and interactions between athletes that characterize successful performance in different sports is an important challenge for all sport practitioners. This book guides the reader in understanding how an ecological dynamics framework for use of artificial intelligence (AI) can be implemented to interpret sport performance and the design of practice contexts. By examining how AI methodologies are utilized in team games, such as football, as well as in individual sports, such as golf and climbing, this book provides a better understanding of the kinematic and physiological indicators that might better capture athletic performance by looking at the current state-of-the-art AI approaches. Artificial Intelligence in Sport Performance Analysis provides an all-encompassing perspective in an innovative approach that signals practical applications for both academics and practitioners in the fields of coaching, sports analysis, and sport science, as well as related subjects such as engineering, computer and data science, and statistics.
Abstract Biomechanical analysis is a powerful tool in the evaluation of movement dysfunction in orthopaedic and neurologic populations. Three-dimensional (3D) motion capture systems are widely used, accurate systems, but are costly and not available in many clinical settings. The Microsoft Kinect™ has the potential to be used as an alternative low-cost motion analysis tool. The purpose of this study was to assess concurrent validity of the Kinect™ with Brekel Kinect software in comparison to Vicon Nexus during sagittal plane gait kinematics. Twenty healthy adults (nine male, 11 female) were tracked while walking and jogging at three velocities on a treadmill. Concurrent hip and knee peak flexion and extension and stride timing measurements were compared between Vicon and Kinect™. Although Kinect measurements were representative of normal gait, the Kinect™ generally under-estimated joint flexion and over-estimated extension. Kinect™ and Vicon hip angular displacement correlation was very low and error was large. Kinect™ knee measurements were somewhat better than hip, but were not consistent enough for clinical assessment. Correlation between Kinect™ and Vicon stride timing was high and error was fairly small. Variability in Kinect™ measurements was smallest at the slowest velocity. The Kinect™ has basic motion capture capabilities and with some minor adjustments will be an acceptable tool to measure stride timing, but sophisticated advances in software and hardware are necessary to improve Kinect™ sensitivity before it can be implemented for clinical use.
Cerebral palsy is leading cause of physical disability in childhood. This interferes with normal ambulation. Rectus femoris muscle spasticity has been reported to be a cause of stiff knee gait in the patients with cerebral palsy. This study aimed to analyse changes in sagittal knee joint kinematic parameters following botulinum toxin injection to the rectus femoris. Fifteen patients were included in the study. A Vicon Motion capture system with Plug-in-Gait model was used for the pre- and post-treatment gait analysis of knee joint angles as well as knee joint flexion velocity. The results from statistical analysis revealed that there were no significant differences in the knee joint range of motion as well as velocity of knee joint flexion in the pre-swing and early swing phase of gait cycle. Although statistical results did not show any significant improvement, it was noted that there was variation in the outcomes for individual patients, with some showing improvements.
Background
The Microsoft Kinect sensor (Kinect) is potentially a low-cost solution for clinical and home-based assessment of movement symptoms in people with Parkinson's disease (PD). The purpose of this study was to establish the accuracy of the Kinect in measuring clinically relevant movements in people with PD.
Methods
Nine people with PD and 10 controls performed a series of movements which were measured concurrently with a Vicon three-dimensional motion analysis system (gold-standard) and the Kinect. The movements included quiet standing, multidirectional reaching and stepping and walking on the spot, and the following items from the Unified Parkinson's Disease Rating Scale: hand clasping, finger tapping, foot, leg agility, chair rising and hand pronation. Outcomes included mean timing and range of motion across movement repetitions.
Results
The Kinect measured timing of movement repetitions very accurately (low bias, 95% limits of agreement <10% of the group mean, ICCs >0.9 and Pearson's r > 0.9). However, the Kinect had varied success measuring spatial characteristics, ranging from excellent for gross movements such as sit-to-stand (ICC = .989) to very poor for fine movement such as hand clasping (ICC = .012). Despite this, results from the Kinect related strongly to those obtained with the Vicon system (Pearson's r > 0.8) for most movements.
Conclusions
The Kinect can accurately measure timing and gross spatial characteristics of clinically relevant movements but not with the same spatial accuracy for smaller movements, such as hand clasping.
Study design:
Cross-sectional laboratory study.
Objectives:
To assess differences in hip strength, iliotibial band length, and hip and knee mechanics during running between male runners with iliotibial band syndrome (ITBS) and healthy controls.
Background:
Flexibility, strength, and running mechanics are commonly assessed in patients with ITBS. However, these variables have not been evaluated concurrently in this population.
Methods:
Thirty-four men participated (17 healthy, 17 ITBS). Hip strength was measured with a handheld dynamometer, and iliotibial band length was assessed using an inclinometer while performing the Ober test. Kinetic and 3-D kinematic data were obtained during running. Kinematic variables of interest included frontal and transverse plane hip and knee joint angles during early stance. Independent-samples t tests, as well as effect sizes, were used to assess group differences.
Results:
Compared to the control group, persons with ITBS had a significantly lower Ober measurement (1.2°), weaker hip external rotators (1.2 Nm/kg), greater hip internal rotation (3.7°), and greater knee adduction (3.6°). However, only hip internal rotation and knee adduction exceeded the minimal detectable difference value.
Conclusion:
Our results suggest that intervention strategies that target neuromuscular control of the hip and knee may be indicated for males with ITBS.
Background
The evaluation of foot posture in a clinical setting is useful to screen for potential injury, however disagreement remains as to which method has the greatest clinical utility. An inexpensive and widely available imaging system, the Microsoft Kinect™, may possess the characteristics to objectively evaluate static foot posture in a clinical setting with high accuracy. The aim of this study was to assess the intra-rater reliability and validity of this system for assessing static foot posture.
Methods
Three measures were used to assess static foot posture; traditional visual observation using the Foot Posture Index (FPI), a 3D motion analysis (3DMA) system and software designed to collect and analyse image and depth data from the Kinect. Spearman’s rho was used to assess intra-rater reliability and concurrent validity of the Kinect to evaluate foot posture, and a linear regression was used to examine the ability of the Kinect to predict total visual FPI score.
Results
The Kinect demonstrated moderate to good intra-rater reliability for four FPI items of foot posture (ρ = 0.62 to 0.78) and moderate to good correlations with the 3DMA system for four items of foot posture (ρ = 0.51 to 0.85). In contrast, intra-rater reliability of visual FPI items was poor to moderate (ρ = 0.17 to 0.63), and correlations with the Kinect and 3DMA systems were poor (absolute ρ = 0.01 to 0.44). Kinect FPI items with moderate to good reliability predicted 61% of the variance in total visual FPI score.
Conclusions
The majority of the foot posture items derived using the Kinect were more reliable than the traditional visual assessment of FPI, and were valid when compared to a 3DMA system. Individual foot posture items recorded using the Kinect were also shown to predict a moderate degree of variance in the total visual FPI score. Combined, these results support the future potential of the Kinect to accurately evaluate static foot posture in a clinical setting.
Dynamic movements of the hand, fingers and thumb are difficult to measure due to the versatility and complexity of movement inherent in function. An innovative approach to measuring hand kinematics is proposed and validated. The proposed system utilises the Microsoft KinectTM and goes beyond gesture recognition, to develop a validated measurement technique of finger kinematics. The proposed system adopted landmark definition (validated through ground truth estimation against assessors) and grip classification algorithms, including kinematic definitions (validated against a laboratory-based motion capture system). The results of the validation show 78% accuracy when identifying specific markerless landmarks. In addition, comparative data with a previously validated kinematic measurement technique show accuracy of MCP±10° (average absolute error (AAE) = 2.4°), PIP±12° (AAE = 4.8°) and DIP±11° (AAE = 4.8°). These results are notably better than clinically based alternative manual measurement techniques. The ability to measure hand movements, and therefore functional dexterity, without interfering with underlying composite movements, is the paramount objective to any bespoke measurement system. The proposed system is the first validated markerless measurement system using the Microsoft KinectTM that is capable of measuring finger joint kinematics. It is suitable for home-based motion capture for the hand and therefore achieves this objective.
Consumer-grade range cameras such as the Kinect sensor have the potential to be used in mapping applications where accuracy requirements are less strict. To realize this potential insight into the geometric quality of the data acquired by the sensor is essential. In this paper we discuss the calibration of the Kinect sensor, and provide an analysis of the accuracy and resolution of its depth data. Based on a mathematical model of depth measurement from disparity a theoretical error analysis is presented, which provides an insight into the factors influencing the accuracy of the data. Experimental results show that the random error of depth measurement increases with increasing distance to the sensor, and ranges from a few millimeters up to about 4 cm at the maximum range of the sensor. The quality of the data is also found to be influenced by the low resolution of the depth measurements.
Over the last decade, digital photography has entered the mainstream. Inexpensive, miniaturized cameras are now routinely included in consumer electronics. Digital projection is poised to make a similar breakthrough, with a variety of vendors offering small, low-cost projectors. As a result, active imaging is a topic of renewed interest in the computer graphics community. In particular, low-cost homemade 3D scanners are now within reach of students and hobbyists with modest budgets. This course provides beginners with the mathematics, software, and practical details they need to leverage projector-camera systems in their own D scanning projects. An example-driven approach is used throughout; each new concept is illustrated using a practical scanner implemented with off-the-shelf parts. The course concludes by detailing how these new approaches are used in rapid prototyping, entertainment, cultural heritage, and web-based applications.
Female runners have a high incidence of developing patellofemoral pain. Abnormal mechanics are thought to be an important contributing factor to patellofemoral pain. However, the contribution of abnormal trunk, hip, and foot mechanics to the development of patellofemoral pain within this cohort remains elusive. Therefore the aim of this study was to determine if significant differences during running exist in hip, trunk and foot kinematics between females with and without patellofemoral pain.
32 female runners (16 patellofemoral pain, 16 healthy control) participated in this study. All individuals underwent an instrumented gait analysis. Between-group comparisons were made for hip adduction, hip internal rotation, contra-lateral pelvic drop, contra-lateral trunk lean, rearfoot eversion, tibial internal rotation, as well as forefoot dorsiflexion and abduction
The patellofemoral pain group had significantly greater peak hip adduction and hip internal rotation. No differences in contra-lateral pelvic drop were found. A trend towards reduced contra-lateral trunk lean was found in the patellofemoral pain group. No significant differences were found in any of the rearfoot or forefoot variables but significantly greater shank internal rotation was found in the patellofemoral pain group.
We found greater hip adduction, hip internal rotation and shank internal rotation in female runners with patellofemoral pain. We also found less contra-lateral trunk lean in the patellofemoral pain group. This may be a potential compensatory mechanism for the poor hip control seen. Rehabilitation programs that correct abnormal hip and shank kinematics are warranted in this population.
The majority of research on gait mechanics in knee osteoarthritis has focused on people with medial compartment involvement. As a result, little is known about the gait mechanics of people with the less common, lateral compartment disease.
The objective of this study was to compare walking mechanics--specifically, differences in frontal-plane lower-extremity kinematics and kinetics--in people with medial knee osteoarthritis, people with lateral knee osteoarthritis, and people who were healthy.
A cross-sectional design was used.
Fifteen people with medial knee osteoarthritis, 15 people with lateral knee osteoarthritis, and 15 people who were healthy (control group) were recruited for the study. All participants underwent a gait analysis at an intentional walking speed. The variables of interest for the study were peak frontal-plane moments and angles and angular excursions of the lower extremity during the stance phase of gait. Data were statistically analyzed with a one-way analysis of variance.
Participants with lateral knee osteoarthritis exhibited significantly less knee adduction excursion, lower peak knee abduction moment, and lower peak rear-foot eversion compared with the control group and the medial knee osteoarthritis group.
Participants in the control group were approximately 10 years younger than participants with knee osteoarthritis. Despite this difference, neither body mass index nor gait speed, each of which is a factor with a stronger influence on gait mechanics, differed among the groups.
Participants with lateral knee osteoarthritis exhibited frontal-plane gait mechanics at the knee and rear foot that were different from those of participants with medial knee osteoarthritis. The results of this study may guide the development of interventions specific to treating people with lateral knee osteoarthritis.
The experimental study of joint kinematics in three dimensions requires the description and measurement of six motion components. An important aspect of any method of description is the ease with which it is communicated to those who use the data. This paper presents a joint coordinate system that provides a simple geometric description of the three-dimensional rotational and translational motion between two rigid bodies. The coordinate system is applied to the knee and related to the commonly used clinical terms for knee joint motion. A convenient characteristic of the coordinate system shared by spatial linkages is that large joint displacements are independent of the order in which the component translations and rotations occur.
The effectiveness of anterior cruciate ligament reconstruction for restoring normal knee kinematics is largely unknown, particularly during sports movements generating large, rapidly applied forces.
Under dynamic in vivo loading, significant differences in 3-dimensional kinematics exist between anterior cruciate ligament-reconstructed knees and the contralateral, uninjured knees.
Prospective, in vivo laboratory study.
Kinematics of anterior cruciate ligament-reconstructed and contralateral (uninjured) knees were evaluated for 6 subjects during downhill running 4 to 12 months after anterior cruciate ligament reconstruction, using a 250 frame/s stereoradiographic system. Anatomical reference axes were determined from computed tomography scans. Kinematic differences between the uninjured and reconstructed limbs were evaluated with a repeated-measures analysis of variance.
Anterior tibial translation was similar for the reconstructed and uninjured limbs. However, reconstructed knees were more externally rotated on average by 3.8 +/- 2.3 degrees across all subjects and time points (P =.0011). Reconstructed knees were also more adducted, by an average of 2.8 +/- 1.6 degrees (P =.0091). Although differences were small, they were consistent in all subjects.
Anterior cruciate ligament reconstruction failed to restore normal rotational knee kinematics during dynamic loading.
Although further study is required, these abnormal motions may contribute to long-term joint degeneration associated with anterior cruciate ligament injury/reconstruction.
Over the centuries the evolution of methods for the capture of human movement has been motivated by the need for new information on the characteristics of normal and pathological human movement. This study was motivated in part by the need of new clinical approaches for the treatment and prevention of diseases that are influenced by subtle changes in the patterns movement. These clinical approaches require new methods to measure accurately patterns of locomotion without the risk of artificial stimulus producing unwanted artifacts that could mask the natural patterns of motion. Most common methods for accurate capture of three-dimensional human movement require a laboratory environment and the attachment of markers or fixtures to the body's segments. These laboratory conditions can cause unknown experimental artifacts. Thus, our understanding of normal and pathological human movement would be enhanced by a method that allows the capture of human movement without the constraint of markers or fixtures placed on the body. In this paper, the need for markerless human motion capture methods is discussed and the advancement of markerless approaches is considered in view of accurate capture of three-dimensional human movement for biomechanical applications. The role of choosing appropriate technical equipment and algorithms for accurate markerless motion capture is critical. The implementation of this new methodology offers the promise for simple, time-efficient, and potentially more meaningful assessments of human movement in research and clinical practice. The feasibility of accurately and precisely measuring 3D human body kinematics for the lower limbs using a markerless motion capture system on the basis of visual hulls is demonstrated.
Human motion capture is frequently used to study musculoskeletal biomechanics and clinical problems, as well as to provide realistic animation for the entertainment industry. The most popular technique for human motion capture uses markers placed on the skin, despite some important drawbacks including the impediment to the motion by the presence of skin markers and relative movement between the skin where the markers are placed and the underlying bone. The latter makes it difficult to estimate the motion of the underlying bone, which is the variable of interest for biomechanical and clinical applications. A model-based markerless motion capture system is presented in this study, which does not require the placement of any markers on the subject's body. The described method is based on visual hull reconstruction and an a priori model of the subject. A custom version of adapted fast simulated annealing has been developed to match the model to the visual hull. The tracking capability and a quantitative validation of the method were evaluated in a virtual environment for a complete gait cycle. The obtained mean errors, for an entire gait cycle, for knee and hip flexion are respectively 1.5 degrees (+/-3.9 degrees ) and 2.0 degrees (+/-3.0 degrees ), while for knee and hip adduction they are respectively 2.0 degrees (+/-2.3 degrees ) and 1.1 degrees (+/-1.7 degrees ). Results for the ankle and shoulder joints are also presented. Experimental results captured in a gait laboratory with a real subject are also shown to demonstrate the effectiveness and potential of the presented method in a clinical environment.
This paper presents a novel variant of structured lighting which exploits the inherent blur in the projector system to overcome the discrepancy in resolution between typical Digital SLR cameras and typical projector systems. More specifically, the scheme estimates the coordinates of the projection of each illuminated scene point in the projector frame with subpixel precision and this additional level of accuracy helps to improve the quality of the resulting 3D reconstructions.
The recent availability of the Kinect™ sensor, a cost-effective markerless motion capture system (MLS), offers interesting possibilities in clinical functional analysis and rehabilitation. However, neither validity nor reproducibility of this device is known yet. These two parameters were evaluated in this study. Forty-eight volunteers performed shoulder abduction, elbow flexion, hip abduction and knee flexion motions; the same protocol was repeated one week later to evaluate reproducibility. Movements were simultaneously recorded by the Kinect (with Microsoft Kinect SDK v.1.5) MLS and a traditional marker-based stereophotogrammetry system (MBS). Considering the MBS as reference, discrepancies between MLS and MBS were evaluated by comparing the range of motion (ROM) between both systems. MLS reproducibility was found to be statistically similar to MBS results for the four exercises. Measured ROMs however were found different between the systems.
The movement of surface mounted targets (SMT) on a shell at the mid-shank and of bone mounted targets attached to the distal shank using a Percutaneous Skeletal Tracker (PST) were simultaneously measured during free-speed walking of three adult subjects having different body types. Surface movement errors in shank kinematic estimates were determined by expressing the segmental motion derived from the SMT relative to the PST-based segment coordinate system (SCS) located at the segment center of gravity. The greatest errors were along and around the shank longitudinal axis, with peak magnitudes of 10 mm of translation and 8° of rotation in one subject. Estimates of knee joint center locations differed by less than 11 mm in each SCS direction. Differences in estimates of net knee joint forces and moments were most prominent during stance phase, with magnitudes up to 39 N in the shank mediolateral direction and 9 N.m about the mediolateral axis. The differences in kinetics were primarily related to the effect of segment position and orientation on the expression of joint forces and on the magnitude and expression of joint moments.
Figure A.1 Walking Trial—Marker Locations and Mass and Frame Rate Information Table A.1 Raw Coordinate Data (cm) Table A.2(a) Filtered Marker Kinematics—Rib Cage and Greater Trochanter (Hip) Table A.2(b) Filtered Marker Kinematics—Femoral Lateral Epicondyle (Knee) and Head of Fibula Table A.2(c) Filtered Marker Kinematics—Lateral Malleolus (Ankle) and Heel Table A.2(d) Filtered Marker Kinematics—Fifth Metatarsal and Toe Table A.3(a) Linear and Angular Kinematics—Foot Table A.3(b) Linear and Angular Kinematics—Leg Table A.3(c) Linear and Angular Kinematics—Thigh Table A.3(d) Linear and Angular Kinematics—½ HAT Table A.4 Relative Joint Angular Kinematics—Ankle, Knee, and Hip Table A.5(a) Reaction Forces and Moments of Force—Ankle and Knee Table A.5(b) Reaction Forces and Moments of Force—Hip Table A.6 Segment Potential, Kinetic, and Total Energies—Foot, Leg, Thigh, and ½ HAT Table A.7 Power Generation/Absorption and Transfer—Ankle, Knee, and Hip
Gait retraining programs are prescribed to assist in the rehabilitation process of many clinical conditions. Using lateral trunk lean modification as the model, the aim of this study was to assess the concurrent validity of kinematic data recorded using a marker-based 3D motion analysis (3DMA) system and a low-cost alternative, the Microsoft Kinect™ (Kinect), during a gait retraining session. Twenty healthy adults were trained to modify their gait to obtain a lateral trunk lean angle of 10°. Real-time biofeedback of the lateral trunk lean angle was provided on a computer screen in front of the subject using data extracted from the Kinect skeletal tracking algorithm. Marker coordinate data were concurrently recorded using the 3DMA system, and the similarity and equivalency of the trunk lean angle data from each system were compared. The lateral trunk lean angle data obtained from the Kinect system without any form of calibration resulted in errors of a high (>2°) magnitude (mean error=3.2±2.2°). Performing global and individualized calibration significantly (P<0.001) improved this error to 1.7±1.5° and 0.8±0.8° respectively. With the addition of a simple calibration the anatomical position coordinates of the Kinect can be used to create a real-time biofeedback system for gait retraining. Given that this system is low-cost, portable and does not require any sensors to be attached to the body, it could provide numerous advantages when compared to laboratory-based gait retraining systems.
Purpose:
Little is known of the potential long-term gait alterations that occur after an anterior cruciate ligament (ACL) reconstruction. In particular, variables, such as impact loading, which have been previously associated with joint deterioration, have not been studied in walking and running after an ACL reconstruction. The purpose of this study was to define the alterations in impact forces, loading rates, and the accompanying sagittal plane kinematic and kinetic mechanics at the time of impact between the ACL-reconstructed group and a healthy control group.
Methods:
Forty females (20 with ACL reconstruction and 20 controls) participated in the study. An instrumented gait analysis was performed on all subjects. Between-group and between-limb comparisons were made for the initial vertical impact force, loading rate, and sagittal plane knee and hip angles as well as moments.
Results:
During walking and running, the ACL cohort had significantly greater initial vertical impact force (P = 0.002 and P = 0.001, respectively) and loading rates (P = 0.03 and P = 0.01, respectively), as well as a smaller knee extensor moment and hip angle during walking (P = 0.000 and P = 0.01, respectively). There was a trend toward a smaller knee moment and hip angle during running (P = 0.08 and P = 0.06, respectively) as well as a larger hip extensor moment during walking (P = 0.06) in the ACL group. No differences were found for hip extensor moment during running and for knee angles between groups during walking or running. Lastly, no between-limb differences were found for any variable.
Conclusions:
Gait deviations such as elevated impact loading and loading rates do not resolve long term after an individual has resumed previous activity levels and these may contribute to the greater risk of early joint degeneration in this population.
Jumping and cutting activities are investigated in many laboratories attempting to better understand the biomechanics associated with non-contact ACL injury. Optical motion capture is widely used; however, it is subject to soft tissue artifact (STA). Biplanar videoradiography offers a unique approach to collecting skeletal motion without STA. The goal of this study was to compare how STA affects the six-degrees-of-freedom motion of the femur and tibia during a jump-cut maneuver associated with non-contact ACL injury. Ten volunteers performed a jump-cut maneuver while their landing leg was imaged using optical motion capture (OMC) and biplanar videoradiography. The within-bone motion differences were compared using anatomical coordinate systems for the femur and tibia, respectively. The knee joint kinematic measurements were compared during two periods: before and after ground contact. Over the entire activity, the within-bone motion differences between the two motion capture techniques were significantly lower for the tibia than the femur for two of the rotational axes (flexion/extension, internal/external) and the origin. The OMC and biplanar videoradiography knee joint kinematics were in best agreement before landing. Kinematic deviations between the two techniques increased significantly after contact. This study provides information on the kinematic discrepancies between OMC and biplanar videoradiography that can be used to optimize methods employing both technologies for studying dynamic in vivo knee kinematics and kinetics during a jump-cut maneuver.
Clinically feasible methods of assessing postural control such as timed standing balance and functional reach tests provide important information, however, they cannot accurately quantify specific postural control mechanisms. The Microsoft Kinect™ system provides real-time anatomical landmark position data in three dimensions (3D), and given that it is inexpensive, portable and simple to setup it may bridge this gap. This study assessed the concurrent validity of the Microsoft Kinect™ against a benchmark reference, a multiple-camera 3D motion analysis system, in 20 healthy subjects during three postural control tests: (i) forward reach, (ii) lateral reach, and (iii) single-leg eyes-closed standing balance. For the reach tests, the outcome measures consisted of distance reached and trunk flexion angle in the sagittal (forward reach) and coronal (lateral reach) planes. For the standing balance test the range and deviation of movement in the anatomical landmark positions for the sternum, pelvis, knee and ankle and the lateral and anterior trunk flexion angle were assessed. The Microsoft Kinect™ and 3D motion analysis systems had comparable inter-trial reliability (ICC difference=0.06±0.05; range, 0.00-0.16) and excellent concurrent validity, with Pearson's r-values >0.90 for the majority of measurements (r=0.96±0.04; range, 0.84-0.99). However, ordinary least products analyses demonstrated proportional biases for some outcome measures associated with the pelvis and sternum. These findings suggest that the Microsoft Kinect™ can validly assess kinematic strategies of postural control. Given the potential benefits it could therefore become a useful tool for assessing postural control in the clinical setting.
Field tests were conducted to assess the clinical performance characteristics of seven optical-based and one electromagnetic-based biomechanical measurement system. A device placed in the center of the calibrated volume enabled the analysis of four optical system properties: (1) the ability to measure the distance between two constantly visible markers rotating in the volume, (2) the ability to measure motion associated with a static marker, (3) the ability to reconstruct position-time histories of markers that were visible to alternating sets of two or three cameras, and (4) the ability to measure the motion of a marker that moved in close proximity to a second marker. Results indicated that five of the seven optical systems produced less than 2.0 mm RMS errors when measuring fully visible moving markers, and typically less that 1.0 mm RMS error when measuring the stationary marker. All passive optical systems confused marker identifications when markers moved within 2 mm of each other in a 3 m long volume. The electromagnetic device was tested by mounting two sensors at a fixed distance and orientation, and measuring their variability as they moved in various patterns within a pre-defined volume. The electromagnetic system produced real-time results, but was clearly susceptible to repeatable interference from metal in the volume.
A novel approach for accurate markerless motion capture combining a precise tracking algorithm with a database of articulated models is presented. The tracking approach employs an articulated iterative closest point algorithm with soft-joint constraints for tracking body segments in visual hull sequences. The database of articulated models is derived from a combination of human shapes and anthropometric data, contains a large variety of models and closely mimics variations found in the human population. The database provides articulated models that closely match the outer appearance of the visual hulls, e.g. matches overall height and volume. This information is paired with a kinematic chain enhanced through anthropometric regression equations. Deviations in the kinematic chain from true joint center locations are compensated by the soft-joint constraints approach. As a result accurate and a more anatomical correct outcome is obtained suitable for biomechanical and clinical applications. Joint kinematics obtained using this approach closely matched joint kinematics obtained from a marker based motion capture system.
The ICP algorithm has been extensively used in computer vision for regis- tration and tracking purposes. The original formulation of this method is restricted to the use of non-articulated models. A straightforward generali- sation to articulated structures is achievable through the joint minimisation of all the structure pose parameters, for example using Levenberg-Marquardt (LM) optimisation. However, in this approach the aligning transformation cannot be estimated in closed form, like in the original ICP, and the approach heavily suffers from local minima. To overcome this limitation, some au- thors have extended the straightforward generalisation at the cost of giving up some of the properties of ICP. In this paper, we present a generalisation of ICP to articulated structures, which preserves all the properties of the original algorithm. The key idea is to divide the articulated body into parts, which can be aligned rigidly in the way of the original ICP, with additional constraints to keep the articulat