Accuracy of classifying the movement strategy in the functional reach test using a markerless motion capture system

The purpose of this study was to examine the accuracy of classifying the movement strategy in the functional reach test (FRT) using a markerless motion capture system (MLS) on the basis of values acquired with a marker-based motion capture system (MBS). Sixty young, injury-free individuals participated in this study. The task action involved reaching forward in the standing position. Using the Microsoft Kinect v2 as an MLS and Vicon as a MBS, the coordinates of the hip joints, knee joints and ankle joints were measured. The hip and ankle joint angles during the task were calculated from the coordinate data. These angles between MLS and MBS were compared using a paired t-test. The accuracy of movement strategy defined using MLS was examined based on the MBS. A t-test showed a significant difference in both the hip and ankle joint angles between systems (p?<?.01). However, in case of using data of left ankle joint, indices of the classification accuracy of MLS were 0.825 for sensitivity, 1.000 for specificity, infinity for positive likelihood ratio and 0.175 for negative likelihood ratio. The results for the right joint angle were similar to those of the left joint angle. Although the absolute measures in the hip and joint angles obtained using MLS differ from MBS, the MLS may be useful for accurately classifying the movement strategy adopted in the FRT.     

Comparison of Marker-Based and Stereo Radiography Knee Kinematics in Activities of Daily Living

Movement of the marker positions relative to the body segments obscures in vivo joint level motion. Alternatively, tracking bones from radiography images can provide precise motion of the bones at the knee but is impracticable for measurement of body segment motion. Consequently, researchers have combined marker-based knee flexion with kinematic splines to approximate the translations and rotations of the tibia relative to the femur. Yet, the accuracy of predicting six degree-of-freedom joint kinematics using kinematic splines has not been evaluated. The objectives of this study were to (1) compare knee kinematics measured with a marker-based motion capture system to kinematics acquired with high speed stereo radiography (HSSR) and describe the accuracy of marker-based motion to improve interpretation of results from these methods, and (2) use HSSR to define and evaluate a new set of knee joint kinematic splines based on the in vivo kinematics of a knee extension activity. Simultaneous measurements were recorded from eight healthy subjects using HSSR and marker-based motion capture. The marker positions were applied to three models of the lower extremity to calculate tibiofemoral kinematics and compared to kinematics acquired with HSSR. As demonstrated by normalized RMSE above 1.0, varus–valgus rotation (1.26), medial–lateral (1.26), anterior–posterior (2.03), and superior–inferior translations (4.39) were not accurately measured. Using kinematic splines improved predictions in varus–valgus (0.81) rotation, and medial–lateral (0.73), anterior–posterior (0.69), and superior–inferior (0.49) translations. Using splines to predict tibiofemoral kinematics as a function knee flexion can lead to improved accuracy over marker-based motion capture alone, however this technique was limited in reproducing subject-specific kinematics.     

Comparative abilities of Microsoft Kinect and Vicon 3D motion capture for gait analysis

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.     

Validity of the total body centre of gravity during gait using a markerless motion capture system

Purpose: The purpose of this study was to examine the validity of total body centre of gravity (COG) measurement during gait with markerless motion capture system (MLS) on the basis of values acquired with a marker-based motion capture system (MBS).

Materials and methods: Thirty young healthy subjects walked on a flat surface as coordinate data from their bodies were acquired using the Kinect v2 (as a MLS) and Vicon systems (as a MBS). COG was calculated using coordinate data of the total body. Comparisons of COG ensemble curves in the mediolateral and vertical directions were performed between MLS and MBS throughout the gait cycle. The relative consistency between these systems was assessed using Pearson correlation coefficients.

Results: The COG trajectory made by using MLS data followed the trend of the COG trajectory with MBS in the mediolateral direction. In the vertical direction, however, the COG trajectories did not match between two systems. High correlation coefficients (r >?0.79) were observed from 30% to 80% of the gait cycle. The greatest difference of COG between MLS and MBS in the mediolateral direction was 1.1?mm. Differences in the vertical direction appeared to be proportional to the distance between the participant and the Kinect v2 sensor.

Conclusion: In the mediolateral direction, COG calculated with MLS data during gait was validated with COG calculated on the basis of a MBS. Further correction of systematic error is necessary to improve the validity of COG calculations in the vertical direction.     

光学表面成像系统实时运动监测精度研究

目的:探究光学表面成像系统实时运动监测的精度。方法:将30例患者的呼吸曲线输入到模体中模拟呼吸运动， 同时利用Catalyst系统对模体进行实时运动监测，比较系统监测的呼吸曲线与参考曲线，从而得到光学表面成像系统实时 运动监测的精度。结果:光学表面成像系统监测的呼吸曲线与参考曲线具有较高的一致性，相关系数均大于0.99，显著相 关。监测误差的平均值为（0.24±0.04）mm，并且随着呼吸信号频率的增加而减小。结论:光学表面成像系统的实时运动 监测精度较高，可用于对患者呼吸运动的监测。在进行呼吸门控治疗时，应考虑呼吸监测系统引入的误差。     

An improved technique for increasing the accuracy of joint-to-ground distance tracking in kinect v2 for foot-off and foot contact detection

The Kinect sensor has been widely used in different applications such as rehabilitation and gait analysis. Whilst Kinect v2 was released with several improvements over its predecessor, it still incorporates depth-map intrinsic inaccuracies. This results in inconsistencies in skeletal-data acquisition, especially in joint localisation and distance-to-ground tracking with respect to the Kinect’s 3-D Cartesian coordinate reference point. This research proposes a correction technique based on the two-point linear equation derived from the information gathered from different subjects’ skeletal data and data regression analysis to compensate the inaccuracies in joint-to-ground data collection. The research also proposes a new footsteps detection method based on skeletal data and plane detection techniques that calculates a footstep by using the ankle’s Euclidean distance from the floor, regardless of the subject’s distance from the camera. The results show that after the correction technique was applied, data acquisition proved to be consistent and more accurate within a distance range of 1.6–2.9 m from the Kinect camera, regardless of the subject’s location to the camera’s reference point. Moreover, the inconsistency of joint data read by the Kinect was reduced from 25.69% to 5.25% and the footsteps detection accuracy increased from 42.85% to 79.76% on average for both legs.     

基于深度学习融合算法的无标记点步态分析系统

目的 以有标记点三维运动捕捉系统(MoCap)为金标准,基于双向长短时记忆(bi-lateral long short term memory, BiLSTM)递归神经网络和线性回归算法构建深度学习融合模型,减小深度传感器的系统误差,从而提高深度传感器下肢运动学分析的准确性。方法 招募10名健康男性大学生进行步态分析,应用MoCap系统和Kinect V2传感器同时采集数据。通过Cleveland Clinic及Kinect逆运动学模型分别计算下肢关节角度。以MoCap系统为目标,Kinect系统得到的角度为输入构建数据集,分别用BiLSTM算法和线性回归算法构建学习模型,得到系统误差修正后的下肢关节角度。使用留一交叉验证法评估模型的性能。采用多重相关系数(coefficient of multiple correlations, CMC)及均方根误差(root mean square error, RMSE)表示下肢关节角度波形曲线相似程度以及平均误差。结果 BiLSTM网络比线性回归算法更能够处理高度非线性的回归问题,尤其是在髋关节内收/外展、髋关节内旋/外旋和踝关节趾屈/背屈角度...     

Marker-based validation of a biplane fluoroscopy system for quantifying foot kinematics

IntroductionRadiostereometric analysis has demonstrated its capacity to track precise motion of the bones within a subject during motion. Existing devices for imaging the body in two planes are often custom built systems; we present here the design and marker-based validation of a system that has been optimized to image the foot during gait.MethodsMechanical modifications were made to paired BV Pulsera C-arms (Philips Medical Systems) to allow unfettered gait through the imaging area. Image quality improvements were obtained with high speed cameras and the correction of image distorting artifacts. To assess the system's accuracy, we placed beads at known locations throughout the imaging field, and used post processing software to calculate their apparent locations.ResultsDistortion correction reduced overall RMS error from 6.56 mm to 0.17 mm. When tracking beads in static images a translational accuracy of 0.094 ± 0.081 mm and rotational accuracy of 0.083 ± 0.068° was determined. In dynamic trials simulating speeds seen during walking, accuracy was 0.126 ± 0.122 mm.DiscussionThe accuracies and precisions found are within the reported ranges from other such systems. With the completion of marker-based validation, we look to model-based validation of the foot during gait.     

Validation of three-dimensional model-based tibio-femoral tracking during running

The purpose of this study was to determine the accuracy of a radiographic model-based tracking technique that measures the three-dimensional in vivo motion of the tibio-femoral joint during running. Tantalum beads were implanted into the femur and tibia of three subjects and computed tomography (CT) scans were acquired after bead implantation. The subjects ran 2.5m/s on a treadmill positioned within a biplane radiographic system while images were acquired at 250 frames per second. Three-dimensional implanted bead locations were determined and used as a "gold standard" to measure the accuracy of the model-based tracking. The model-based tracking technique optimized the correlation between the radiographs acquired via the biplane X-ray system and digitally reconstructed radiographs created from the volume-rendered CT model. Accuracy was defined in terms of measurement system bias, precision and root-mean-squared (rms) error. Results were reported in terms of individual bone tracking and in terms of clinically relevant tibio-femoral joint translations and rotations (joint kinematics). Accuracy for joint kinematics was as follows: model-based tracking measured static joint orientation with a precision of 0.2 degrees or better, and static joint position with a precision of 0.2mm or better. Model-based tracking precision for dynamic joint rotation was 0.9+/-0.3 degrees , 0.6+/-0.3 degrees , and 0.3+/-0.1 degrees for flexion-extension, external-internal rotation, and ab-adduction, respectively. Model-based tracking precision when measuring dynamic joint translation was 0.3+/-0.1mm, 0.4+/-0.2mm, and 0.7+/-0.2mm in the medial-lateral, proximal-distal, and anterior-posterior direction, respectively. The combination of high-speed biplane radiography and volumetric model-based tracking achieves excellent accuracy during in vivo, dynamic knee motion without the necessity for invasive bead implantation.     

Development and assessment of a Microsoft Kinect based system for imaging the breast in three dimensions

Three-dimensional surface imaging technologies have been used in the planning and evaluation of breast reconstructive and cosmetic surgery. The aim of this study was to develop a 3D surface imaging system based on the Microsoft Kinect and assess the accuracy and repeatability with which the system could image the breast. A system comprising two Kinects, calibrated to provide a complete 3D image of the mannequin was developed. Digital measurements of Euclidean and surface distances between landmarks showed acceptable agreement with manual measurements. The mean differences for Euclidean and surface distances were 1.9 mm and 2.2 mm, respectively. The system also demonstrated good intra- and inter-rater reliability (ICCs > 0.999). The Kinect-based 3D surface imaging system offers a low-cost, readily accessible alternative to more expensive, commercially available systems, which have had limited clinical use.     

Evaluation of suitability of a micro-processing unit of motion analysis for upper limb tracking

The aim of this study is to assess the suitability of a micro-processing unit of motion analysis (MPUMA), for monitoring, reproducing, and tracking upper limb movements. The MPUMA is based on an inertial measurement unit, a 16-bit digital signal controller and a customized algorithm. To validate the performance of the system, simultaneous recordings of the angular trajectory were performed with a video-based motion analysis system. A test of the flexo-extension of the shoulder joint during the active elevation in a complete range of 120º of the upper limb was carried out in 10 healthy volunteers. Additional tests were carried out to assess MPUMA performance during upper limb tracking. The first, a 3D motion reconstruction of three movements of the shoulder joint (flexo-extension, abduction–adduction, horizontal internal–external rotation), and the second, an upper limb tracking online during the execution of three movements of the shoulder joint followed by a continuous random movement without any restrictions by using a virtual model and a mechatronic device of the shoulder joint. Experimental results demonstrated that the MPUMA measured joint angles that are close to those from a motion-capture system with orientation RMS errors less than 3º.     

Multileaf collimator tracking integrated with a novel x-ray imaging system and external surrogate monitoring

We have previously developed a tumour tracking system, which adapts the aperture of a Siemens 160 MLC to electromagnetically monitored target motion. In this study, we exploit the use of a novel linac-mounted kilovoltage x-ray imaging system for MLC tracking. The unique in-line geometry of the imaging system allows the detection of target motion perpendicular to the treatment beam (i.e. the directions usually featuring steep dose gradients). We utilized the imaging system either alone or in combination with an external surrogate monitoring system. We equipped a Siemens ARTISTE linac with two flat panel detectors, one directly underneath the linac head for motion monitoring and the other underneath the patient couch for geometric tracking accuracy assessments. A programmable phantom with an embedded metal marker reproduced three patient breathing traces. For MLC tracking based on x-ray imaging alone, marker position was detected at a frame rate of 7.1 Hz. For the combined external and internal motion monitoring system, a total of only 85 x-ray images were acquired prior to or in between the delivery of ten segments of an IMRT beam. External motion was monitored with a potentiometer. A correlation model between external and internal motion was established. The real-time component of the MLC tracking procedure then relied solely on the correlation model estimations of internal motion based on the external signal. Geometric tracking accuracies were 0.6 mm (1.1 mm) and 1.8 mm (1.6 mm) in directions perpendicular and parallel to the leaf travel direction for the x-ray-only (the combined external and internal) motion monitoring system in spite of a total system latency of ~0.62 s (~0.51 s). Dosimetric accuracy for a highly modulated IMRT beam--assessed through radiographic film dosimetry--improved substantially when tracking was applied, but depended strongly on the respective geometric tracking accuracy. In conclusion, we have for the first time integrated MLC tracking with x-ray imaging in the in-line geometry and demonstrated highly accurate respiratory motion tracking.     

Development and Validation of a Motion and Loading System for a Rat Knee Joint <Emphasis Type="Italic">In Vivo</Emphasis>

The influence of biomechanical stimuli on modulating cartilage homeostasis is well recognized. However, many aspects of cellular mechanotransduction in cartilage remain unknown. We developed a computer-controlled joint motion and loading system (JMLS) to study the biological response of cartilage under well-characterized mechanical loading environments. The JMLS was capable of controlling (i) angular displacement, (ii) motion frequency, (iii) magnitude of the axial compressive load applied to the moving joint, and it featured real-time monitoring. The accuracy and repeatability of angular position measurements, the kinematic misalignment error as well as the repositioning error of the JMLS were evaluated. The effectiveness of the JMLS in implementing well-defined loading protocols such as moderate Passive Motion Loading (PML) and increased Compressive Motion Loading (CML) were tested. The JMLS demonstrated remarkable accuracy and reliability for the measurement and kinematics tests. Moreover, the effectiveness test demonstrated the ability of the JMLS to produce an effective stimulus via PML that led to the suppression of the catabolic effects of immobilization. Interestingly, the biological response of the CML group was catabolic and exhibited a pattern similar to that observed in the immobilization group. This novel non-invasive system may be useful for joint biomechanics studies that require different treatment conditions of load and motion in vivo.     

基于Kinect体感交互技术的上肢关节活动度测量方法

为了简化人体上肢关节活动度测量过程,提高测量效率和准确度,实现活动度的自动测量,提出一种基于Kinect体感交互技术的测量方法。通过Kinect传感器,捕捉上肢各关节点位置,将关节点数据传至计算机;通过检测上肢体运动平面偏离人体冠状面、矢状面和轴向面的距离,规范上肢的测量动作;自动计算肩、肘等关节各科目的活动度数值,由体感人机交互界面实时反馈测量结果。用该方法对40位受测者分5组(每组8人)进行测量试验,并根据t检验法,得到肩关节外展、内收、屈曲、伸展以及肘关节屈曲科目的|t|值,分别为1.53、1.17、1.30、1.58和1.47,均小于2.776,测量结果与期望值无显著差异,表明系统的测量精确度较高。     

Kinect人体运动捕捉误差及其空间分布

目的 针对微软发布的体感捕捉设备Kinect开展精度测试实验,检测系统定位的位置误差及其空间分布,为基于Kinect体感交互的生物医学工程应用提供参考.方法 本实验利用Kinect和高精度运动捕捉设备NDI Optotrak同时进行人体运动捕捉,并采用Kinect和NDI进行测量,结果取相对位移值进行比较.以NDI的测量数据作为真实数据的有效近似,评估Kinect的测量精度.结果 本实验最终测得Kinect识别精度误差为（0.0 283±0.0 186）m,均方根误差为0.0 303 m,水平角度识别范围为51.49＆#176;.结论 Kinect的精度和稳定性在cm级,可应用于康复训练、手术室设备控制等生物医学工程领域,但对于精度要求更高的领域,如智能手术机器人的控制等,其精度和稳定性还有待提高.     

Development of a real-time three-dimensional spinal motion measurement system for clinical practice

This paper presents an inertial based sensing system for real-time three-dimensional measurement of human spinal motion, in a portable and non-invasive manner. Applications of the proposed system range from diagnosis of spine injury to postural monitoring, on-field as well as in the lab setting. The system is comprised of three inertial measurement sensors, respectively attached and calibrated to the head, torso and hips, based on the subject’s anatomical planes. Sensor output is transformed into meaningful clinical parameters of rotation (twist), flexion-extension and lateral bending of each body segment, with respect to calibrated global reference space. Modeling the spine as a compound flexible pole model allows dynamic measurement of three-dimensional spine motion, which can be animated and monitored in real-time using our interactive GUI. The accuracy of the proposed sensing system has been verified with subject trials using a VICON optical motion measurement system. Experimental results indicate an error of less than 3.1° in segment orientation tracking.     

Tracking evoked responses to auditory and visual stimuli in fetuses exposed to maternal high‐risk conditions

Magnetoencephalography (MEG) has been successfully applied to record fetal auditory (auditory evoked response [AER]) and visual evoked responses (VER). In this study, we report the AER and VER development trajectory by tracking the evoked response detectability and latency from recordings starting at 27 weeks of gestation in pregnancies classified as high risk. Fetal MEG and ultrasound recordings were performed on 158 pregnant women, and the total number of fetal auditory and visual tests conducted was 321 and 237, respectively. The overall evoked response analysis showed 237 AER (73.8%) and 164 VER detections (69.2%). The mean AER latency was 290.7 (SD 125.5) ms and the mean VER latency was 293.7 (SD 114.5) ms. The rate of decrease (95% confidence limits) in average AER and VER first‐peak latency between 100–350 ms was 1.97 (?1.86, +5.81) ms/week and 1.35 (?3.83, +6.53) ms/week, respectively. This trend in high‐risk fetuses conforms to the general trajectory of decrease in latency with gestational age progression, even though this decrease was non‐significant, as reported in the case of normal growing fetuses. Although there was a significant difference in detection rates between male and female fetuses, this was not reflected in either latency values or the sensory modality applied. Furthermore, the main factors that had the most significant effect on response detectability included the presence of intervening layers of adipose tissue between the fetal head and stimulus source and an increase in the maternal body mass index.     

Synchronized tumour tracking with electromagnetic transponders and kV x-ray imaging: evaluation based on a thorax phantom

Intrafractional organ motion remains a source of error in conformal radiotherapy of dynamic targets such as tumours of the lung or of the prostate. The purpose of this work was to devise a method for the continuous and routine measurement of intrafractional organ motion. The method consists of a combination of an electromagnetic (EM), internal marker-based tracking system with the on-board kilovoltage x-ray imaging system of a modern treatment machine. The EM system continuously tracks the target, while x-ray images can be acquired simultaneously if demand arises. An image processing algorithm has been developed to automatically localize and track the EM markers in the x-ray images. We have demonstrated simultaneous target tracking using the EM system and x-ray imaging of a mobile target inside a programmable thorax phantom. The target motion was very well reproduced by both systems. The comparability of the target locations reported by both systems was established (better than 0.25 mm up to target velocities of 3 cm s(-1)). One immediate use of the synchronized system was shown: the generation of a 4D cone beam computed tomography data set using the EM system for the measurement of motion. In conclusion, we have developed a system for the routine measurement of intrafractional motion that continuously provides the 3D position of the target with the ability to acquire images of the treatment field only when needed, thereby eliminating avoidable imaging dose to the patient.     

The accuracy and repeatability of an automatic 2D-3D fluoroscopic image-model registration technique for determining shoulder joint kinematics

Fluoroscopic imaging, using single plane or dual plane images, has grown in popularity to measure dynamic in vivo human shoulder joint kinematics. However, no study has quantified the difference in spatial positional accuracy between single and dual plane image-model registration applied to the shoulder joint. In this paper, an automatic 2D-3D image-model registration technique was validated for accuracy and repeatability with single and dual plane fluoroscopic images. Accuracy was assessed in a cadaver model, kinematics found using the automatic registration technique were compared to those found using radiostereometric analysis. The in vivo repeatability of the automatic registration technique was assessed during the dynamic abduction motion of four human subjects. The in vitro data indicated that the error in spatial positional accuracy of the humerus and the scapula was less than 0.30mm in translation and less than 0.58° in rotation using dual plane images. Single plane accuracy was satisfactory for in-plane motion variables, but out-of-plane motion variables on average were approximately 8 times less accurate. The in vivo test indicated that the repeatability of the automatic 2D-3D image-model registration was 0.50mm in translation and 1.04° in rotation using dual images. For a single plane technique, the repeatability was 3.31mm in translation and 2.46° in rotation for measuring shoulder joint kinematics. The data demonstrate that accurate and repeatable shoulder joint kinematics can be obtained using dual plane fluoroscopic images with an automatic 2D-3D image-model registration technique; and that out-of-plane motion variables are less accurate than in-plane motion variables using a single plane technique.     

Validity of the verbal IQ as a short form of the Wechsler adult intelligence scale-revised

Assessed the validity of the Verbal IQ as a short form of the WAIS-R. Ss were 104 psychiatric patients with means for age, education, and Full Scale IQ of 36.22 (SD = 9.04), 12.46 (SD = 1.98), and 93.94 (SD = 12.19), respectively. A correlation of 0.93 (p < 0.001) between the Verbal and Full Scale IQs was found. The average Verbal IQ exceeded the average Full Scale IQ by a small (i. e., 1.65 IQ points) but statistically significant amount (p < 0.001). Thirty-three (32%) Ss showed changes in their intelligence categories when the Verbal IQ was compared to the Full Scale IQ. However, when the Verbal IQ was banded by the standard error of measurement (SE_M = ± 3) and the precision range was compared to the Full Scale IQ, results indicated 88% agreement. If clinicians must rely on the Verbal IQ as an estimate of the Full Scale, reporting the score in conjunction with a precision range will increase its accuracy.