No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Gait phase classification for in-home gait assessment
... Related range sensor and homevideo based systems, which cost about £700, such as [9][10][11], that build on the work of [12], with Pro-Trainer motion analysis software (Sports Motion, Inc., Cardiff, CA), offer gait analysis outside the gait laboratory, e.g. in local clinics and at homes. Similar to other range sensor and home-video based gait analysis systems [2,[13][14][15][16][17][18][19][20][21] and inertial measurement unit (IMU) based gait analysis systems [22][23][24][25][26], the gait parameters obtained after data processing can be sent to physiatrists for clinical consultation, indicating the potential for tele-rehabilitation [27][28][29][30][31]. It is shown in [32] that a 2D video tracker software provides similar accuracy to VICON 3D system for knee angle measurement but not for measurement of the ankle angle over time. ...
... Our system addresses some of the drawbacks of related range sensor and home-video based systems [2,9,11,[13][14][15][16][17][18][19]41] and IMU systems [22][23][24][25][26] namely: (i) unlike [9], there are no colour restrictions on the background or the participant's clothing; (ii) in contrast to [9], which is validated on only one healthy volunteer with one walking trial with no gold standard benchmark, we validate our proposed system's knee angle against the gold standard VICON MX Giganet 6xT40 and 6xT160 (VICON Motion Systems Ltd., Oxford, UK, approximately £ 250,000) optical motion analysis system (the same gold standard as used by Ugbolue et al. [11]); (iii) unlike systems of [11] and Pro-Trainer and Siliconcoach (Siliconcoach Ltd., Dunedin, New Zealand) as used by the authors in [42,43] that require significant manual effort, our system autonomously tracks the markers attached to the joints and calculates the knee angle; the only operational effort required is for marker-template selection for tracking initialisation which is done via a user-friendly graphical user interface (GUI); (iv) unlike the passive marker system [41] that is only validated on one side of the body without any benchmarking systems, our system is validated on both sides of the body with a gold standard VICON optical motion analysis system; (v) 3D Kinect range sensor-based systems [13-18, 20, 21] cannot reliably capture relatively fast body motion, since Kinect operates at only 30 frames per second (fps), whereas our system operates at 210 fps; (vi) like other range sensor and home-video based systems, our system is non-intrusive to the participants, which is in contrast to state-of-the-art IMU gait analysis systems [22][23][24][25][26]. However, with only a 2D camera in our gait analysis system, its drawback lies in the following two aspects: (i) estimation of the human joint locations using our system is less accurate compared to 3D Kinect-based range sensor systems, and (ii) the gait parameters derived from the 2D images in our system are less reliable than those derived from the inertial data in IMU systems. ...
... Our future work will be focused on further improvement of performance and potential measurement capability of more gait parameters using a stereo 2D-camera system or a single depth sensing device [2,[13][14][15][16][17][18][19] with a high frame rate, without sacrificing portability, to remove the parallax error, and leverage the 3D information for quantifying a larger number of gait parameters such as hip, knee, and ankle angles in both the sagittal and frontal planes, and pelvis tilt, calculating temporal-spatial parameters (step length and width, stride length, step time, cadence and step length symmetry), gait speed and measuring sagittal/ frontal plane knee motion, but at an increased processing complexity. Furthermore, neural-network-based methods are what we could consider in future research that involves evaluations with public availability of large labelled datasets of walking trials. ...
While optical motion analysis systems can provide high-fidelity gait parameters, they are usually impractical for local
clinics and home use, due to high cost, requirement for large space, and lack of portability. In this study, we focus on a cost-effective
and portable, single-camera gait analysis solution, based on video acquisition with calibration, autonomous detection of frames-ofinterest,
Kalman-filter+Structural-Similarity-based marker tracking, and autonomous knee angle calculation. The proposed system
is tested using 15 participants, including 10 stroke patients and 5 healthy volunteers. The evaluation of autonomous frames-ofinterest
detection shows only 0.2% difference between the frame number of the detected frame compared to the frame number of
the manually labelled ground truth frame, and thus can replace manual labelling. The system is validated against a gold standard
optical motion analysis system, using knee angle accuracy as metric of assessment. The accuracy investigation between the RGBand
the grayscale-video marker tracking schemes shows that the grayscale system suffers from negligible accuracy loss with a
significant processing speed advantage. Experimental results demonstrate that the proposed system can automatically estimate
the knee angle, with R-squared value larger than 0.95 and Bland-Altman plot results smaller than 3.0127 degrees mean error.
... It is to compare and find optimal alignment between two given time-sequences. Furthermore, Ye et al [60] utilizes DTW to compute a distance matrix for gait pattern extraction while Su et al [61] applying DTW measure the similarity of joint data between "at home exercises "and" in hospital exercises". ...
... Su et al then evaluate the performance by using Adaptive Neuro-Fuzzy Inference System (ANFIS) which integrates a neural network and a fuzzy logic [61]. Nomm et al practiced Neural Network based model, NN-based ANARX (Additive Nonlinear Auto Regressive exogenous), in their monitoring system as it can adjust the system according to the specific needs of each patient [62] whereas Ye et al used NN-based on nonlinear autoregressive with exogenous (NARX) for gait phase classification and Enhanced Random Decision Forest (ERF) for missing features cases [60]. Next, Nahavandi et al trained a Random Decision Forest (RDF) for generalising a learning model in order to discriminate between seven RULA-scored sets of postures [63]. ...
Kinect-based physical rehabilitation grows significantly as a mechanism for clinical assessment and rehabilitation due to its flexibility, low-cost and markerless system for human action capture. It is also an approach to provide convenience for for patients’ exercises continuation at home. In this paper, we discuss a review of the present Kinect-based physiotherapy and assessment for rehabilitation patients to provide an outline of the state of art, limitation and issues of concern as well as suggestion for future work in this approach. The paper is constructed into three main parts. The introduction was discussed on physiotherapy exercises and the limitation of current Kinect-based applications. Next, we also discuss on Kinect Skeleton Joint and Kinect Depth Map features that being used widely nowadays. A concise summary with significant findings of each paper had been tabulate for each feature; Skeleton Joints and Depth Map. Afterwards, we assemble a quite number of classification method that being implemented for activity recognition in past few years. © 2018 Institute of Advanced Engineering and Science. All rights reserved.
... Consequently, running gait assessment has moved towards instrumented approaches such as wearable technology including inertial measurement units (IMU), force/pressure plate analysis [8][9][10], as well as three-dimensional (3D) motion tracking [11,12] in an effort to provide reliable, reproducible outcomes. Despite research-grade wearable technology's utility in providing a wide range of gait outcomes, they are currently limited in use due to the cost [13], tethering of peripheral technologies, and reliance upon bespoke environments with expert assistance [9]. For example, wearable sensors or 3D-motion reflective markers often rely upon precise anatomical placement for optimal use [14,15]. ...
Running gait assessment is essential for development of technique optimization strategies as well as to inform injury prevention and rehabilitation. Currently, running gait assessment relies on (i) visual assessment, exhibiting subjectivity and limited reliability, or (ii) use of instrumented approaches, which often carry high costs and can be intrusive due to attachment of equipment to the body. Here use of an IoT-enabled markerless computer vision smartphone application based upon Googles pose estimation model BlazePose was evaluated for running gait assess-ment for use in low-resource settings. That human pose estimation architecture was used to ex-tract contact time, swing time, step time, knee flexion angle and foot strike location from a large cohort of runners. The gold-standard Vicon 3D motion capture system was used as a reference. The proposed approach performs robustly, demonstrating good (ICC(2,1) > 0.75) to excellent (ICC(2,1) > 0.90) agreement in all running gait outcomes. Additionally, temporal outcomes exhibit low mean error (0.01-0.014s) in left foot outcomes. However, there are some discrepancies in right foot outcomes, due to occlusion. This study demonstrates that the proposed low-cost and markerless system provides accurate running gait assessment outcomes. The approach may help routine running gait assessment in low-resource environments.
... Publication [47] presents a comparison between different alghorithms to extract gait phases from inertial units worn by the subject and surface electromyography, while [48] presents inertial units and an adaptive Bayesian approach for recognition of walking activities. Reference [49] presents instead a classification method based on a single 3D depth camera and 12 key feature, to provide a reliable and low cost way to assess the gait phases. All the sensors used in these works, however, are either worn by the user, or monitor a limited volume of space where the user can walk. ...
The paper presents a novel method for the classification of gait phases for power gait orthosis users based on machine learning. The classification uses depth images collected from a Time of Flight camera embedded in the crutches employed for the assisted gait. The machine learning algorithm foresees an initial phase of data collection and processing, identifying the 3D points belonging to the foot and those belonging to the floor. From these, a feature set is computed analyzing the values of percentiles of distances of the foot from the floor, and passed to a modified version of Random Forest classifier, called Sigma-z Random Forest. The classifier considers the uncertainties associated to each feature set and provides both the classification of the gait phase (stance or swing) and an associated confidence value. In this work, we propose the use of the confidence value to improve the reliability of the gait phase classification, by applying an optimized threshold to the confidence value obtained for each new frame. The algorithm has been tested on different subjects and environments. An average classification accuracy of 87.3% has been obtained (+6.3% with respect to the standard random forest classifier), with a minor loss of unclassifiable frames. Results highlight that unclassifiable samples are usually associated to transitions between stance and swing.
... Recent research on WSS for biomedical applications can be divided into two major areas. One area focuses on recognition of daily activities, such as walking feature recognition [2], [3], walking condition classification [1], [4]- [6] and gait phase detection [7]- [10], in which the kinematic data obtained from inertial sensors (accelerometers or gyroscopes) are directly used as inputs to inference techniques. The second area focuses on accurate measurement of human motion data such as joint angles, or 3D body segment positions and orientations. ...
There are many important challenges in gait analysis, which has many applications in healthcare, rehabilitation, therapy, and exercise training. However, gait analysis is typically performed in a gait laboratory, which is inaccessible to the general population and is not available in natural gait environments (e.g., outdoors). In this paper, we discuss the development of a high-fidelity, cost-effective, wireless sensor network to address the challenge of efficient gait monitoring in real-world walking scenarios. The sensor network is designed in a modular way to capture plantar forces and knee angle, angular velocity, and angular acceleration. A force module called a smart insole is designed to measure the plantar forces. The module is comprised of force sensitive resistors (FSRs) and a signal conditioning circuit. Various signal conditioning techniques, including a novel technique called transfrequency, are investigated to provide a linear mapping for FSR measurements and to provide data acquisition fidelity. The motion module includes a low-cost inertial measurement unit (IMU) augmented with a Kalman filter to provide filtered knee kinematics. A qualitative evaluation of the sensor network communication module is achieved by considering the internal communication protocols between the modules and the external wireless transmission protocol used to deliver data to an end-point terminal PC. Experiments are conducted to validate the motion and force modules. Then, the overall, integrated system is compared to gold standard laboratory results, demonstrating a successful application for gait identification. The results show that the sensor network accurately captures important gait parameters and features.
... Inertial sensors are also extensively used for ADL analysis, balance assessment and gait analysis of older people . The use of a Kinect sensor in gait analysis has gained attention of researchers for gait pattern recognition and classification [24,25]. ...
Background:
Frailty assessment is a critical approach in assessing the health status of older people. The clinical tools deployed by geriatricians to assess frailty can be grouped into two categories; using a questionnaire-based method or analyzing the physical performance of the subject. In performance analysis, the time taken by a subject to complete a physical task such as walking over a specific distance, typically three meters, is measured. The questionnaire-based method is subjective, and the time-based performance analysis does not necessarily identify the kinematic characteristics of motion and their root causes. However, kinematic characteristics are crucial in measuring the degree of frailty.
Results:
The studies reviewed in this paper indicate that the quantitative analysis of activity of daily living, balance and gait are significant methods for assessing frailty in older people. Kinematic parameters (such as gait speed) and sensor-derived parameters are also strong markers of frailty. Seventeen gait parameters are found to be sensitive for discriminating various frailty levels. Gait velocity is the most significant parameter. Short term monitoring of daily activities is a more significant method for frailty assessment than is long term monitoring and can be implemented easily using clinical tests such as sit to stand or stand to sit. The risk of fall can be considered an outcome of frailty.
Conclusion:
Frailty is a multi-dimensional phenomenon that is defined by various domains; physical, social, psychological and environmental. The physical domain has proven to be essential in the objective determination of the degree of frailty in older people. The deployment of inertial sensor in clinical tests is an effective method for the objective assessment of frailty.
Segmentation of a gait cycle is an important issue to perform sub-phase wise assessment. It is performed by identifying some crucial gait events. Generally, manual annotation is followed in clinics to detect events. But, it is laborious and prone to error. On contrary, machine intelligence–based automatic event detection systems are gradually becoming popular. However, most of such systems use high-end sensors which make it expensive. A low-cost alternative solution is of high need in this domain.
nside clinical research, gait analysis is a fundamental part of the functional evaluation of the human body's movement. Its evaluation has been carried out through different methods and tools, which allow early diagnosis of diseases, and monitoring and assessing the effectiveness of therapeutic plans applied to patients for rehabilitation. The observational method is one of the most used in specialized centers in Colombia; however, to avoid any possible errors associated with the subjectivity observation, technological tools that provide quantitative data can support this method. This paper deals with the methodological process for developing a computational tool and hardware device for the analysis of gait, specifically on articular kinematics of the knee. This work develops a prototype based on the fusion of inertial measurement units (IMU) data as an alternative for the attenuation of errors associated with each of these technologies. A videogrammetry technique measured the same human gait patterns to validate the proposed system, in terms of accuracy and repeatability of the recorded data. Results showed that the developed prototype successfully captured the knee-joint angles of the flexion-extension motions with high consistency and accuracy in with the measurements obtained from the videogrammetry technique. Statistical analysis (ICC and RMSE) exhibited a high correlation between the two systems for the measures of the joint angles. These results suggest the possibility of using an IMU-based prototype in realistic scenarios for accurately tracking a patient’s knee-joint kinematics during a human gait.
Rapid advances in cost-effective and non-invasive depth sensors, and the development of reliable and real-time 3D skeletal data estimation algorithms, have opened up a new application area in computer vision - statistical analysis of human kinematic data for fast, automated assessment of body movements. These assessments can play important roles in sports, medical diagnosis, physical therapy, elderly monitoring and related applications. This paper develops a comprehensive geometric framework for quantification and statistical evaluation of kinematic features. The key idea is to avoid analysis of individual joints, as is the current paradigm, and represent movements as temporal evolutions, or trajectories, on shape space of full body skeletons. This allows metrics with appropriate invariance properties to be imposed on these trajectories and leads to definitions of higher-level features, such as
$spatial\ symmetry\ (sS),\ temporal\ symmetry\ (tS),\ action's\ velocity (Vl)$
and
$ body's\ balance\ (Bl)$
, during performance of an action. These features exploit skeletal symmetries in space and time, and capture motion cadence to naturally quantify motions of individual subjects. The study of these features as functional data allows us to formulate certain hypothesis tests in feature space. This, in turn, leads to validation of existing assumptions and discoveries of new relationships between kinematics and demographic factors, such as age, gender, and athletic training. We use the clinically validated K3Da kinect dataset to illustrate these ideas, and hope these tools will lead to discovery of new relationships between full body kinematic features and demographic, health, and wellness factors that are clinically relevant.
Recent advances in image acquisition and analysis have resulted in disruptive innovation in physical rehabilitation systems facilitating cost-effective, portable, video-based gait assessment. While these inexpensive motion capture systems, suitable for home rehabilitation, do not generally provide accurate kinematics measurements on their own, image processing algorithms ensure gait analysis that is accurate enough for rehabilitation programs. This paper proposes high-accuracy classification of gait phases and muscle actions, using readings from low-cost motion capture systems. First, 12 gait parameters, drawn from the medical literature, are defined to characterize gait patterns. These proposed parameters are then used as input to our proposed
multi-channel time-series classification
and gait phase reconstruction methods. Proposed methods fully utilize temporal information of gait parameters, thus improving the final classification accuracy. The validation, conducted using 126 experiments, with 6 healthy volunteers and 9 stroke survivors with manually-labelled gait phases, achieves state-of-art classification accuracy of gait phase with lower computational complexity compared to previous solutions.
<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sup
Gait Initiation Failure (GIF) is one of the most disabling gait disturbances seen in advanced Parkinson’s disease (PD). Gait Initiation is a complex motor task that requires motor and cognitive processing to enable the correct selection, timing and scaling of movement. Failure to initiate the first step often precipitates falls and leads to significant morbidity. However, the brain mechanisms underlying GIF remain unknown. This study utilized an ambulatory electroencephalography (EEG) technique to investigate the brain dynamic changes underlying GIF and aims to detect the occurrence of GIF in four PD patients. We sought to determine whether episodes of GIF might be associated with a characteristic brain signal that could be detected by surface EEG. This preliminary investigation analyzed the EEG signals through power spectra density (PSD) and centroid frequency (CF) to show that the GIF episodes were associated with significant increases in the high beta band (21-38Hz) across the central, frontal, occipital and parietal EEG sites. By implementing PSD and CF as input features with two-layer Back Propagation neural networks as a classifier, the proposed system was able to detect GIF events with a classification performance of 84.27% sensitivity and 84.80% accuracy. This is the first study to show cortical dynamic changes associated with GIF in Parkinson’s disease, providing valuable information to enhance the performance of future GIF detection that could be translated into clinical practice.
Gait impairment is a prevalent and important difficulty for patients with multiple sclerosis (MS), a common neurological disorder. An easy to use tool to objectively evaluate gait in MS patients in a clinical setting can assist clinicians to perform an objective assessment. The overall objective of this study is to develop a framework to quantify gait abnormalities in MS patients using the Microsoft Kinect for the Windows sensor; an inexpensive, easy to use, portable camera. Specifically, we aim to evaluate its feasibility for utilization in a clinical setting, assess its reliability, evaluate the validity of gait indices obtained, and evaluate a novel set of gait indices based on the concept of dynamic time warping. In this study, ten ambulatory MS patients, and ten age and sex-matched normal controls were studied at one session in a clinical setting with gait assessment using a Kinect camera. The expanded disability status scale (EDSS) clinical ambulation score was calculated for the MS subjects, and patients completed the Multiple Sclerosis walking scale (MSWS). Based on this study, we established the potential feasibility of using a Microsoft Kinect camera in a clinical setting. Seven out of the eight gait indices obtained using the proposed method were reliable with intraclass correlation coefficients ranging from 0.61 to 0.99. All eight MS gait indices were significantly different from those of the controls (p-values less than 0.05). Finally, seven out of the eight MS gait indices were correlated with the objective and subjective gait measures (Pearson's correlation coefficients greater than 0.40). This study shows that the Kinect camera is an easy to use tool to assess gait in MS patients in a clinical setting.
Clustering governs huge data by organizing similar data objects into groups. Density-based clustering permits composition of data objects on basis of their density distribution. DBSCAN, an illustrious and prominent density-based clustering algorithm gives birth to arbitrary-framed clusters, without requiring preexisting acquaintances on the number of clusters to be produced. The inputs to DBSCAN principal are: dataset required to be mined, radius of neighborhood—Eps (ε), minimum number of points needed to build a cluster (MinPts). DBSCAN clustering desires these two parameters to be given as input manually and automatic detection of these parameters is a very tedious exercise and has a significant influence on clustering result. In this paper, we contemplated a new and efficient density-based clustering algorithm (E-DBSCAN). The consolidated notion of the proposed approach is that it avoids manual intervention of input values. Experimental results demonstrate effectiveness and efficiency of the proposed algorithm on varied domain of datasets.
Stroke is a worldwide healthcare problem, which often causes long-term motor impairment, handicap, and disability. Optical motion analysis systems are commonly used for impairment assessment due to high accuracy. However, the requirement of equipment-heavy and large laboratory space together with operational expertise makes these systems impractical for local clinic and home use. We propose an alternative, cost-effective and portable, decision support system for optical motion analysis, using a single camera. The system relies on detecting and tracking markers attached to subject's joints, data analytics for calculating relevant rehabilitation parameters, visualization, and robust classification based on graph-based signal processing. Experimental results show that the proposed decision support system has the potential to offer stroke survivors and clinicians an alternative, affordable, accurate, and convenient impairment assessment option suitable for home healthcare and telerehabilitation.
Laboratory-based nonwearable motion analysis systems have significantly advanced with robust objective measurement of the limb motion, resulting in quantified, standardized, and reliable outcome measures compared with traditional, semisubjective, observational gait analysis. However, the requirement for large laboratory space and operational expertise makes these systems impractical for gait analysis at local clinics and homes. In this paper, we focus on autonomous gait event detection with our bespoke, relatively inexpensive, and portable, single-camera gait kinematics analysis system. Our proposed system includes video acquisition with camera calibration, Kalman filter + Structural-Similarity-based marker tracking, autonomous knee angle calculation, video-frame-identification-based autonomous gait event detection, and result visualization. The only operational effort required is the marker-template selection for tracking initialization, aided by an easy-to-use graphic user interface. The knee angle validation on 10 stroke patients and 5 healthy volunteers against a gold standard optical motion analysis system indicates very good agreement. The autonomous gait event detection shows high detection rates for all gait events. Experimental results demonstrate that the proposed system can automatically measure the knee angle and detect gait events with good accuracy and thus offer an alternative, cost-effective, and convenient solution for clinical gait kinematics analysis.
An exact classification of different gait phases is essential to enable the control of exoskeleton robots and detect the intentions of users. We propose a gait phase classification method based on neural networks using sensor signals from lower limb exoskeleton robots. In such robots, foot sensors with force sensing registers are commonly used to classify gait phases. We describe classifiers that use the orientation of each lower limb segment and the angular velocities of the joints to output the current gait phase. Experiments to obtain the input signals and desired outputs for the learning and validation process are conducted, and two neural network methods (a multilayer perceptron and nonlinear autoregressive with external inputs (NARX)) are used to develop an optimal classifier. Offline and online evaluations using four criteria are used to compare the performance of the classifiers. The proposed NARX-based method exhibits sufficiently good performance to replace foot sensors as a means of classifying gait phases.
In this paper, we propose a method for temporal segmentation of human repetitive actions based on frequency analysis of kinematic parameters, zero-velocity crossing detection, and adaptive k-means clustering. Since the human motion data may be captured with different modalities which have different temporal sampling rate and accuracy (e.g., optical motion capture systems vs. Microsoft Kinect), we first apply a generic full-body kinematic model with an unscented Kalman filter to convert the motion data into a unified
representation that is robust to noise. Furthermore, we extract the most representative kinematic parameters via the primary frequency analysis. The sequences are segmented based on zero-velocity crossing of the selected parameters followed by an adaptive k-means clustering to identify the repetition segments. Experimental results demonstrate that for the motion data captured by both the motion capture system and the Microsoft Kinect, our proposed algorithm obtains robust segmentation of repetitive action sequences.
In this work, we decided to apply a hierarchical weighted decision, proposed and used in other research fields, for the recognition of gait phases. The developed and validated novel distributed classifier is based on hierarchical weighted decision from outputs of scalar Hidden Markov Models (HMM) applied to angular velocities of foot, shank, and thigh. The angular velocities of ten healthy subjects were acquired via three
uni-axial gyroscopes embedded in inertial measurement units (IMUs) during one walking task, repeated three times, on a treadmill. After validating the novel distributed classifier and scalar and vectorial classifiers-already proposed in the literature, with a cross-validation, classifiers were compared for sensitivity, specificity, and computational load for all combinations of the three targeted anatomical segments. Moreover, the performance of the novel distributed classifier in the estimation of gait variability in terms of mean time and
coefficient of variation was evaluated. The highest values of specificity and sensitivity (>0.98) for the three classifiers examined here were obtained when the angular velocity of the foot was processed. Distributed and vectorial classifiers reached acceptable values (>0.95) when the angular velocity of shank and thigh were analyzed. Distributed and scalar classifiers showed values of computational load about 100 times lower than the one obtained with the vectorial classifier. In addition, distributed classifiers showed an excellent reliability for the evaluation of mean time and a good/excellent reliability for the coefficient of variation. In conclusion, due to the better performance and the small value of computational load, the here proposed novel distributed classifier can be implemented in the real-time application of gait phases recognition, such as to evaluate gait variability in patients or to control active orthoses for the recovery of mobility of lower limb joints.
Microsoft Kinect uses a built-in RGB-D sensor and the skeleton tracking algorithm to capture 3-D movements of the human body. It also has the potential for assessing postural stability, which is fundamental for most motor activities. The aim of this paper was to investigate whether standing balance can be evaluated reliably and validly by this low-cost device. Nine healthy subjects were required to maintain balance during three standing positions (double limb stance with feet apart, double limb stance with feet together and single limb stance). The center of mass (COM) was calculated from the body's kinematic data acquired by the Kinect system and Optotrak Certus motion capture system. The position variability and average velocity of the COM in the horizontal plane were calculated and used to evaluate the subject's balance. These COM parameters from the two systems showed excellent and comparable test-retest reliability (intraclass correlation coefficient $({rm ICC}){>}{0.75}$). In addition, although the average velocity of the COM calculated from Kinect was significantly lower, each COM parameter showed excellent concurrent validity $({rm ICC}{>}0.88)$ and a significant linear relationship $(p<0.001,r>0.930)$ existed between the two systems, which meant that biases may be corrected using linear calibration equations. Therefore, Kinect may be a valid, reliable, and convenient device for assessing standing balance when its measured COM parameters are properly calibrated.
We describe two new approaches to human pose estimation. Both can quickly and accurately predict the 3D positions of body joints from a single depth image without using any temporal information. The key to both approaches is the use of a large, realistic, and highly varied synthetic set of training images. This allows us to learn models that are largely invariant to factors such as pose, body shape, field-of-view cropping, and clothing. Our first approach employs an intermediate body parts representation, designed so that an accurate per-pixel classification of the parts will localize the joints of the body. The second approach instead directly regresses the positions of body joints. By using simple depth pixel comparison features and parallelizable decision forests, both approaches can run super-real time on consumer hardware. Our evaluation investigates many aspects of our methods, and compares the approaches to each other and to the state of the art. Results on silhouettes suggest broader applicability to other imaging modalities.
The use of virtual reality for the provision of motor-cognitive gait training has been shown to be effective for a variety of patient populations. The interaction between the user and the virtual environment is achieved by tracking the motion of the body parts and replicating it in the virtual environment in real time. In this paper, we present the validation of a novel method for tracking foot position and orientation in real time, based on the Microsoft Kinect technology, to be used for gait training combined with virtual reality. The validation of the motion tracking method was performed by comparing the tracking performance of the new system against a stereo-photogrammetric system used as gold standard. Foot position errors were in the order of a few millimeters (average RMSD from 4.9 mm to 12.1 mm in the medio-lateral and vertical directions, from 19.4 mm to 26.5 mm in the anterior-posterior direction); the foot orientation errors were also small (average %RMSD from 5.6% to 8.8% in the medio-lateral and vertical directions, from 15.5% to 18.6% in the anterior-posterior direction). The results suggest that the proposed method can be effectively used to track feet motion in virtual reality and treadmill-based gait training programs.
Mining sequential data is an old topic that has been revived in the last decade, due to the increasing availability of sequential datasets. Most works in this field are centred on the definition and use of a distance (or, at least, a similarity measure) between sequences of elements. A measure called dynamic time warping (DTW) seems to be currently the most relevant for a large panel of applications. This article is about the use of DTW in data mining algorithms, and focuses on the computation of an average of a set of sequences. Averaging is an essential tool for the analysis of data. For example, the K-means clustering algorithm repeatedly computes such an average, and needs to provide a description of the clusters it forms. Averaging is here a crucial step, which must be sound in order to make algorithms work accurately. When dealing with sequences, especially when sequences are compared with DTW, averaging is not a trivial task.Starting with existing techniques developed around DTW, the article suggests an analysis framework to classify averaging techniques. It then proceeds to study the two major questions lifted by the framework. First, we develop a global technique for averaging a set of sequences. This technique is original in that it avoids using iterative pairwise averaging. It is thus insensitive to ordering effects. Second, we describe a new strategy to reduce the length of the resulting average sequence. This has a favourable impact on performance, but also on the relevance of the results. Both aspects are evaluated on standard datasets, and the evaluation shows that they compare favourably with existing methods. The article ends by describing the use of averaging in clustering. The last section also introduces a new application domain, namely the analysis of satellite image time series, where data mining techniques provide an original approach.
With increasing importance given to telerehabilitation, there is a growing need for accurate, low-cost, and portable motion capture systems that do not require specialist assessment venues. This paper proposes a novel framework for motion capture using only a single depth camera, which is portable and cost effective compared to most industry-standard optical systems, without compromising on accuracy. Novel signal processing and computer vision algorithms are proposed to determine motion patterns of interest from infrared and depth data. In order to demonstrate the proposed framework’s suitability for rehabilitation, we developed a gait analysis application that depends on the underlying motion capture
sub-system. Each subject’s individual kinematics parameters, which are unique to that subject, are calculated and these are stored for monitoring individual progress of the clinical therapy. Experiments were conducted on 14 different subjects, 5 healthy and 9 stroke survivors. The results show very close agreement of the resulting relevant joint angles with a 12-camera based VICON system, a mean error of at most 1.75% in detecting gait events w.r.t the manually generated ground-truth, and significant performance improvements in terms of accuracy and execution time compared to a previous Kinect-based system.
Spasticity is a common disorder of the skeletal muscle with a high incidence in industrialised countries. A quantitative measure of spasticity using body-worn sensors is important in order to assess rehabilitative motor training and to adjust the rehabilitative therapy accordingly. We present a new approach to spasticity detection using the Integrated Posture and Activity NEtwork by Medit Aachen (IPANEMA) body sensor network (BSN). For this, a new electromyography (EMG) sensor node was developed and employed in human locomotion. Following an analysis of the clinical gait data of patients with unilateral cerebral palsy, a novel algorithm was developed based on the idea to detect co-activation of antagonistic muscle groups as observed in the exaggerated stretch reflex with associated joint rigidity. The algorithm applies a cross-correlation function to the EMG signals of two antagonistically working muscles and subsequent weighting using a Blackman window. The result is a co-activation index which is also weighted by the signal equivalent energy to exclude positive detection of inactive muscles. Our experimental study indicates good performance in the detection of co-active muscles associated with spasticity from clinical data as well as measurements from a BSN in qualitative comparison with the Modified Ashworth Scale (MAS) as classified by clinical experts. Possible applications of the new algorithm include (but are not limited to) use in robotic sensorimotor therapy to reduce the effect of spasticity.
BACKGROUND, gait event detection is important for diagnosis and evaluation. This is a challenging endeavor due to subjectivity, high amount of data, among other problems. ANFIS, ARX, OE, NARX and models based on NN were developed in order to detect gait events without the problems mentioned. OBJECTIVE, to compare developed models’ performance and determinate the most suitable model for gait events detection.
METHODS, knee joint angle, heel foot switch and toe foot switch during normal walking in a treadmill were collected from a healthy volunteer. Gait events were classified by three experts in human motion. Experts’ mean classification was obtained and all models were trained and tested with the collected data and experts’ mean classification. Fit percentage was obtained to evaluate models performance. RESULTS, Fit percentage was: ANFIS: 79.49%, ARX: 68.8%, OE: 71.39%, NARX: 88.59%, NNARX: 67.66%, NNRARX: 68.25% and NNARMAX: 54.71%.DISCUSSION, NARX had the best performance for gait events classification. For ARX and OE, previous filtering is needed. NN’s models showed the best performance for high frequency components. ANFIS and NARX were able to integrate criteria from three experts for gait analysis.
CONCLUSION, NARX and ANFIS are suitable for gait event identification. Test with additional subjects is needed.
This study aims to validate a commercially available inertial sensor based motion capture system, Xsens MVN BIOMECH using its native protocols, against a camera-based motion capture system for the measurement of joint angular kinematics. Performance was evaluated by comparing waveform similarity using range of motion, mean error and a new formulation of the coefficient of multiple correlation (CMC). Three dimensional joint angles of the lower limbs were determined for ten healthy subjects while they performed three daily activities: level walking, stair ascent, and stair descent. Under all three walking conditions, the Xsens system most accurately determined the flexion/extension joint angle (CMC > 0.96) for all joints. The joint angle measurements associated with the other two joint axes had lower correlation including complex CMC values. The poor correlation in the other two joint axes is most likely due to differences in the anatomical frame definition of limb segments used by the Xsens and Optotrak systems. Implementation of a protocol to align these two systems is necessary when comparing joint angle waveforms measured by the Xsens and other motion capture systems.
Tradition gait analysis systems capture the image of a walking subject from either front view or side view. Since the walking direction allowed by the systems is highly restricted, they are inconvenient for long-term evaluation in casual environments (such as home). This study proposes a human gait analysis system with much less restriction on walking direction. In the system, we use the images obtained from multi-viewing angles and performs human gait analysis based on a set of chosen features, including the center of gravity (COG) and pace length, obtained from human silhouette images. The system successfully extracts gait features from various walking directions and integrates the features obtained from two cameras having orthogonal views. The intergration principle is discussed in details. The integrated feature is then compared with the ideal gait feature obtained from the camera whose viewing direction is perpendicular to the walking path, resulting in very high correlation. This study shows that a vision-based gait analyzer with two orthogonally arranged cameras has the potential to remove the walking direction restriction.
This paper presents a walking pattern classification and a walking distance estimation algorithm using gait phase information. A gait phase information retrieval algorithm was developed to analyze the duration of the phases in a gait cycle (i.e., stance, push-off, swing, and heel-strike phases). Based on the gait phase information, a decision tree based on the relations between gait phases was constructed for classifying three different walking patterns (level walking, walking upstairs, and walking downstairs). Gait phase information was also used for developing a walking distance estimation algorithm. The walking distance estimation algorithm consists of the processes of step count and step length estimation. The proposed walking pattern classification and walking distance estimation algorithm have been validated by a series of experiments. The accuracy of the proposed walking pattern classification was 98.87%, 95.45%, and 95.00% for level walking, walking upstairs, and walking downstairs, respectively. The accuracy of the proposed walking distance estimation algorithm was 96.42% over a walking distance.
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.
In this paper, a novel robust markerless image processing system capable of extracting gait features which can be used for gait analysis is presented. The presented system can deal with images of persons captured in natural indoor scenes. The system's robustness against external influences and different person appearance is achieved by employing the idea of improving the image processing robustness by including feedback control at the image segmentation level. The effectiveness of the proposed system is demonstrated by the comparison of gait features, namely knee angles, extracted automatically with the features directly measured using a goniometer. Also a small database is created to extract the gait pattern of healthy subjects. The obtained data is compared to data from medical literature and is also compared to data obtained from persons having a pathological gait.
This study evaluated a modified, timed version of the "Get-Up and Go" Test (Mathias et al, 1986) in 60 patients referred to a Geriatric Day Hospital (mean age 79.5 years). The patient is observed and timed while he rises from an arm chair, walks 3 meters, turns, walks back, and sits down again. The results indicate that the time score is (1) reliable (inter-rater and intra-rater); (2) correlates well with log-transformed scores on the Berg Balance Scale (r = -0.81), gait speed (r = -0.61) and Barthel Index of ADL (r = -0.78); and (3) appears to predict the patient's ability to go outside alone safely. These data suggest that the timed "Up & Go" test is a reliable and valid test for quantifying functional mobility that may also be useful in following clinical change over time. The test is quick, requires no special equipment or training, and is easily included as part of the routine medical examination.
Kinect-based gait analysis for automatic frailty syndrome assessment
- E Gianaria
Gait assessment using the kinect rgb-d sensor
- J Zhao
Fusion of gait and ecg for biometric user authentication
- M Derawi