Fig 1 - available via license: CC BY
Content may be subject to copyright.
Frames from color video and depth data collected using a depth-sensing camera during structured mobility testing in the community setting. Times (t) are computed based on frame number and indicate the time elapsed since Frame A, the right leg toe-off. In the depth frames, colors code for distance of the object in that pixel from the camera. The green stick figure superimposed on the depth data represents the skeleton of the body pose estimated using motion capture software. The red-highlighted segments in Frame E represent the pelvis, thigh, and shank segments, from which we extracted hip and knee angles. The individual in this figure gave written informed consent (as outlined in the PLOS consent form) to publish this series of video frames.
Source publication
Background
Currently, it is not feasible to obtain laboratory-based measures of joint motion in large numbers of older adults. We assessed the utility of a portable depth-sensing camera for quantifying hip and knee joint motion of older adults during mobility testing in the community.
Methods
Participants were 52 older adults enrolled in the Rush...
Contexts in source publication
Context 1
... portable, depth-sensing video camera originally developed for video gaming (Kinect for Windows, Microsoft Corporation, Redmond, WA) was employed to capture participants' biomechanics during gait. Using infrared light, this device records "depth" video, which quantifies an object's distance from the camera at 30 frames per second (Fig 1). The camera was positioned as shown in Fig 2. Its recording was controlled via software (iPi Recorder version 3.1.1.34, ...
Context 2
... camera was positioned as shown in Fig 2. Its recording was controlled via software (iPi Recorder version 3.1.1.34, iPi Soft, Moscow, Russia) running on a notebook computer (Pavilion 13 x360 Convertible PC, HP, Palo Alto, CA) as participants traversed a walking path (Fig 1). We transferred this data to a workstation (Z620, HP) to run frame-by-frame body pose estimation software (iPi Mocap Studio version 3.1.2.177, iPi Soft) (Fig 1). ...
Context 3
... Soft, Moscow, Russia) running on a notebook computer (Pavilion 13 x360 Convertible PC, HP, Palo Alto, CA) as participants traversed a walking path (Fig 1). We transferred this data to a workstation (Z620, HP) to run frame-by-frame body pose estimation software (iPi Mocap Studio version 3.1.2.177, iPi Soft) (Fig 1). In order to compare hip and knee motion captured using the depthsensing camera with that obtained using a state-of-the-art optoelectronic motion capture system (Qualysis, Gothenborg, Sweden), we first enrolled 10 participants who were able to travel Frames from color video and depth data collected using a depth-sensing camera during structured mobility testing in the community setting. ...
Similar publications
In the last few years, the importance of measuring gait characteristics has increased tenfold due to their direct relationship with various neurological diseases. As patients suffering from Parkinson’s disease (PD) are more prone to a movement disorder, the quantification of gait characteristics helps in personalizing the treatment. The wearable se...
Gait disorders and falls are common in elders and in many clinical conditions, yet they are typically infrequently and subjectively evaluated, limiting prevention and intervention. Completion-time of the Timed-Up-and-Go (TUG) test is a well-accepted clinical biomarker for rating mobility and prediction of falls risk. Using smartphones’ integral acc...
Wearable sensors have the potential to facilitate remote monitoring for patients recovering from knee replacement surgery. Using IMU sensors attached to the patients’ leg, knee flexion can be monitored while the patients are recovering in their home environment. Ideally, these flexion angle measurements will have an accuracy and repeatability at le...
We present the use of a single inertial measurement unit (IMU) worn on the thigh to produce stride-by-stride estimates of walking speed and its spatiotemporal determinants (i.e., stride time and stride length). Ten healthy and eight post-stroke individuals completed a 6-min walk test with an 18-camera motion capture system used for ground truth mea...
Enhancement of activity is one major topic related to the aging society. Therefore, it is necessary to understand people's motion and identify possible risk factors during activity. Technology can be used to monitor motion patterns during daily life. Especially the use of artificial intelligence combined with wearable sensors can simplify measureme...
Citations
... Additionally, the 6-DOF ranges of motion calculated are in close keeping with those outcomes measured by the biplane uoroscopy [25][26][27], which is the gold standard in kinematic analysis. The accuracies of various portable simpli ed motion capture systems are progressing rapidly in recently years [28][29][30], although they cannot yet substitute the biplanar uoroscopy, but the future is worth looking forward to. ...
Background: Anterior cruciate ligament (ACL) injuries are common, but traditional examinations primarily focus on static assessments, lacking the ability to evaluate dynamic knee stability. Therefore, a dynamic knee function scoring system is needed.
Purpose: This study aimed to propose a dynamic scoring system based on a large sample of normative six-degrees-of-freedom (6-DOF) gait kinematics and validate its correlation with conventional outcome measurements when assessing ACL-injured knees.
Methods: A total of 500 healthy Chinese participants were recruited to establish a large dataset. 83 patients with isolated unilateral ACL injury were included for preoperative assessments. An infrared navigation three-dimensional portable knee motion analysis system was used to collect data on the 6-DOF of both knees while walking on a treadmill. A novel 6-DOF scoring system was developed based on the large sample data using the dynamic time warping algorithm. The Spearman correlation coefficients (ρ) were calculated between the 6-DOF score and the Lysholm score, 2000 International Knee Documentation Committee (IKDC) subjective score, and Tegner’s activity scale.
Results: The 95% confidence intervals (CI) for abduction/adduction, internal/external rotation, flexion/extension, anterior/posterior translation, medial/lateral translation, and proximal/distal translation in the 500 healthy participants were as follows: 10.07 ± 4.04°, 15.13 ± 4.85°, 60.56 ± 6.07°, 1.79 ± 0.75 cm, 1.09 ± 0.42 cm, and 1.58 ± 0.54 cm, respectively. The Spearman correlation coefficients (ρ) between the 6-DOF score and the assessment scores were as follows: ρ = 0.375 (p < 0.001) for the Lysholm score, ρ = 0.273 (p = 0.016) for Tegner’s activity scale, and ρ = 0.145 (p = 0.208) for the 2000 IKDC subjective score.
Conclusion: Significant correlations were found between the 6-DOF score and the Lysholm score as well as Tegner’s activity scale. Therefore, the 6-DOF score has the potential to provide comprehensive and meaningful information on the dynamic knee function and stability of ACL-injured patients.
... An unknown potential source for error besides the depth measurement is also the implemented artificial intelligence of the Body racking SDK. Additionally, it is possible that white or black clothes as well as the surrounding infrared light can affect the pose estimation performance [37]. We did not control these circumstances. ...
Background: Accidental falls are a major health issue in older people. One significant and potentially modifiable risk factor is reduced gait stability. Clinicians do not have sophisticated kinematic options to measure this risk factor with simple and affordable systems. Depth-imaging with AI-pose estimation can be used for gait analysis in young healthy adults. However, is it applicable for measuring gait in older adults at a risk of falling? Methods: In this methodological comparison 59 older adults with and without a history of falls walked on a treadmill while their gait pattern was recorded with multiple inertial measurement units and with an Azure Kinect depth-camera. Spatiotemporal gait parameters of both systems were compared for convergent validity and with a Bland-Altman plot. Results: Correlation between systems for stride length (r=.992, p<0.001) and stride time (r=0.914, p<0.001) was high. Bland-Altman plots revealed a moderate agreement in stride length (-0.74 ± 3.68 cm; [-7.96 cm to 6.47 cm]) and stride time (-3.7±54 ms; [-109 ms to 102 ms]). Conclusion: Gait parameters in older adults with and without a history of falls can be measured with inertial measurement units and Azure Kinect cameras. Affordable and small depth-cameras agree with IMUs for gait analysis in older adults with and without an increased risk of falling. However, tolerable accuracy is limited to the average over multiple steps of spatiotemporal parameters derived from the initial foot contact.
... ogy, its real-world applications are often limited to a lab setting. On the other hand, low-cost D-Mocap has been applied for some real-world applications, such as gait assessment [8,15,28,33], rehabilitation [34], human mobility analysis [22], and exercise systems [6]. With human motion enhancement and denoising of the low-quality D-Mocap data, depth sensors could become an inexpensive and versatile alternative for clinical applications, Although there are ample datasets that provide clean high quality Mocap data, there are very few that provide low quality D-Mocap motion data with a high quality counterpart time synced for performance evaluation. ...
... The joint positions of the enhanced human motion data are then evaluated on a joint by joint basis using the euclidean distance from the MHAD Mocap data averaged over all frames (8). These joint-by-joint values are given in Table 2 for the original D-Mocap data and all four enhancement methods. ...
... During the past decade, low cost RGB-D depth sensors have emerged as a promising alternative for motion capture (D-Mocap). They have proven useful in the clinical setting for gait assessment [17], [25], [40], [46], rehabilitation [47], human mobility analysis [32], and exercise systems [14]. In addition, there is a wealth of off-the-shelf or open source software which generates D-Mocap (Fig. 1). ...
... Likewise, the results on the right hand of simulated data drop-out, those of the left knee in the OSU D-Mocap data, and those of the right elbow in the MHAD data are shown in (Figs. 15,16,and 17) respectively. In addition, longer frame sequences are shown in (Figs. ...
We present a novel approach to enhance the quality of human motion data collected by low-cost depth sensors, namely D-Mocap, which suffers from low accuracy and poor stability due to occlusion, interference, and algorithmic limitations. Our approach takes advantage of a large set of high-quality and diverse Mocap data by learning a general motion manifold via the convolutional autoencoder. In addition, the Tobit Kalman filter (TKF) is used to capture the kinematics of each body joint and handle censored measurement distribution. The TKF is incorporated with the autoencoder via latent space optimization, maintaining adherence to the motion manifold while preserving the kinematic nature of the original motion data. Furthermore, due to the lack of an open source benchmark dataset for this research, we have developed an extension of the Berkeley Multimodal Human Action Database (MHAD) by generating D-Mocap data from RGB-D images. The newly extended MHAD dataset is skeleton-matched and time-synced to the corresponding Mocap data and is publicly available. Along with simulated D-Mocap data generated from the CMU Mocap dataset and our self-collected D-Mocap dataset, the proposed algorithm is thoroughly evaluated and compared with different settings. Experimental results show that our approach can improve the accuracy of joint positions and angles as well as skeletal bone lengths by over 50%.
... L OW-cost RGB-D or depth sensors have recently gained prominence in a broad variety of applications ranging from computer vision and computer graphics to entertainment [1] and health care [2]. Particularly, depth sensors have been used to study human gaits as a diagnostic, measuring, and predictive tool. ...
... Gait analysis has historically been performed using expensive optical motion capture (Mocap) systems, which are inefficient in a clinical environment. On the other hand, as a more affordable and practical alternative, depthbased Mocap (D-Mocap) has some advantages that supports markerless motion capture and can be easily set up in a normal setting [1], [3]. However, current D-Mocap has a few major technical limitations that hinders its health care related applications, including poor precision, a restricted sensing range [4], distortion due to self-occlusion [5], and potential sensor interference [6]. ...
In this paper, we propose a novel technique for human motion denoising by jointly optimizing kinematic and anthropometric constraints for a noisy skeleton data. Specifically, we are focused on depth-sensor-based motion capture (D-Mocap) data that are often prone to error, outliers and distortion. To capture human kinematics, we first propose a joint-level Tobit particle filter (TPF) that incorporates a unique observation model to characterize the censored measurement of D-Mocap data. A skeleton-level Differential Evolution (DE) algorithm is then integrated with the sequential Monte Carlo sampling in the TPF, allowing joint-level particles to be re-distributed and re-weighted according to the stability and consistency of skeletal bone lengths as well as the suitability of joint kinematics. This leads to an integrated TPF-DE algorithm that significantly improves the quality of D-Mocap data by making 3D joint trajectories more kinematically admissible and anthropometrically stable. Experimental results on both simulated and real-world D-Mocap show that the errors of joint positions and the bone lengths have been reduced by 30-60%, and the accuracy of joint angles has been improved by 40-60%. The proposed TPF-DE method outperforms the recent filtering-based and deep learning methods and demonstrate the synergy between the TPF and DE algorithms for effective human motion enhancement.
... Synopsizing, although Golden standard systems, for clinical and lab use, achieve sub-millimeter accuracy, yet, reports from labs with such systems still contain significant measurement errors, mainly due to personnel, protocols and less-than-ideal conditions. Furthermore, there is increased interest in new low-cost and portable gait analysis systems [15] which are now forming a new market trend. These systems are typically benchmarked with the aid of the golden standard systems [16]. ...
Scientific gait analysis methods aim to offer objective measurements, to assist physicians towards an accurate diagnosis or pre-diagnosis of ailments before they actually manifest through noticeable symptoms. This paper reviews selected gait analysis system technologies, trends, applications and discusses errors and precision in spatial and angular readings. Furthermore, we propose a novel test and calibration method using a biomimetic rig. To illustrate this, we conduct three tests on an optical single-camera gait analysis system based on a mobile android smart-phone with specially developed software.
... Synopsizing, although Golden standard systems, for clinical and lab use, achieve sub-millimeter accuracy, yet, reports from labs with such systems still contain significant measurement errors, mainly due to personnel, protocols and less-than-ideal conditions. Furthermore, there is increased interest in new low-cost and portable gait analysis systems [15] which are now forming a new market trend. These systems are typically benchmarked with the aid of the golden standard systems [16]. ...
... There are numerous works in the fields of surveillance and entertainment where depth sensors are used for human motion capture (Mocap). Particularly, in medicine and health care, depth sensors have been used for gait analysis as an assessment, diagnostic, or even as a predictive tool [1][2][3][4][5]. Traditionally kinematic gait measurements are obtained by optical Mocap systems, which are considered the gold standard but are impractical in a clinical setting. ...
... IMU sensors have gained popularity in rehabilitation and other motion tracking applications [6,7]. On the other hand, as a more affordable and practical alternative, depth-based Mocap (D-Mocap) has some advantages for markerless motion capture in a normal setting [5,8]. However, D-Mocap suffers from low accuracy due to several technical limitations: the limited imaging range [9], self-occlusions during motion [10], and possible interference [11]. ...
... It is executed by keeping a population of candidate solutions and creating new candidate solutions by mutation and crossover operations between exciting candidates, where the candidate solution with the best fitness will be kept. To reduce the bone length variation, a DE algorithm is employed to optimize the joint position 5 EAI Endorsed Transactions on Bioengineering and Bioinformatics 03 2021 -08 2021 | Volume 1 | Issue 3 | e1 m th bone, its bone-length can be defined by: ...
INTRODUCTION: Recently, low cost RGB-D depth sensors have emerged as a promising alternative for motion capture for clinical gait assessment. However, depth sensor based Mocap (D-Mocap) suffers from low accuracy and poor stability for 3D joint estimation due to noise, self-occlusion, interference, and other technical limitations, which prevents it from being widely used in health-related applications.
... One of the most relevant works found in the area of information gathering on biomechanical movements belongs to Dawe's team [9]. The accuracy and feasibility of using an optical system based on depth detection is tested, quantifying the movement of the hip and knee joints of older adults. ...
A proposed measurement protocol for the lower limbs movement analysis during walking is presented, with the use of a measurement system based on inertial-magnetic motion processing units and an optical system. Initially, the state of the art in terms of methods and tools for the biomechanical capture of movements is shown, to finally explore the protocols used in the health sciences for the gait analysis. The measurement proposal made in this document uses robust features of inertial-magnetic and optical technology that can be used in medical diagnosis. The application of this proposal can generate tools that have a positive impact in the fields of health and medicine.
... Many have attempted to use virtual reality to mimic natural environments, but this is not quite the same as being at a real intersection where unstable gait can actually be quite dangerous (Banducci et al., 2016;Chaddock et al., 2012;Nagamatsu et al., 2011;Neider et al., 2011;Souza Silva et al., 2019 (Matthis et al., 2018). Cheaper methods involve the use of depth sensors, but these techniques only work over small capture volumes (Dawe et al., 2019). The simplest approach is to use dedicated accelerometers to identify step parameters directly from acceleration data, but these devices are still considered specialty hardware (Samà et al., 2011). ...
Physical activity (PA) levels have been declining worldwide, despite the physical health benefits of increased PA having been known and messaged for decades. One possible reason for this is that physical health improvements are too long term, and it may be more effective to promote short-term benefits instead. PA has been shown to improve cognitive function over relatively short timescales and promoting these cognitive benefits may be a better approach for increasing PA levels in the population. What remains unclear is how certain moderators can affect the PA-cognition relationship, and how we can better measure PA in real-world environments. Over three studies, my dissertation aims to examine the effect of moderators like PA intensity and PA time window, as well as develop a novel methodology for measuring low-intensity physical activities, like walking, in outdoor settings. In the first study, I found that PA measured over the week prior to cognitive testing is not predictive of cognitive function, and that this does not depend on PA intensity. In the second study, I found that PA measured over the day prior to testing is associated with improved cognitive performance. In the third study, I examined the bidirectional nature of this relationship, and found that changes in cognition can have a marked impact on outdoor walking behaviour, and that these changes in gait dynamics can be detected using a smartphone-based measurement technique. Overall, my research shows that while PA does appear to be associated with cognitive function, the effects are quite small and dependent on factors like when the PA is measured relative to cognitive testing. In terms of PA promotion, more research needs to be conducted on the limits of this relationship before we can draw any definitive conclusions about the cognitive benefits of PA, as small statistically significant effects don’t necessarily translate to useful practical benefits in the real world.
