Accurate upper body rehabilitation system using kinect
01 Aug 2016-Vol. 2016, pp 4605-4609
TL;DR: An optimization method that utilizes Kinect Depth and RGB information to search for the joint center location that satisfies constraints on body segment length and as well as orientation is proposed, enabling to more accurate measurements for upper limb exercises.
Abstract: The growing importance of Kinect as a tool for clinical assessment and rehabilitation is due to its portability, low cost and markerless system for human motion capture. However, the accuracy of Kinect in measuring three-dimensional body joint center locations often fails to meet clinical standards of accuracy when compared to marker-based motion capture systems such as Vicon. The length of the body segment connecting any two joints, measured as the distance between three-dimensional Kinect skeleton joint coordinates, has been observed to vary with time. The orientation of the line connecting adjoining Kinect skeletal coordinates has also been seen to differ from the actual orientation of the physical body segment. Hence we have proposed an optimization method that utilizes Kinect Depth and RGB information to search for the joint center location that satisfies constraints on body segment length and as well as orientation. An experimental study have been carried out on ten healthy participants performing upper body range of motion exercises. The results report 72% reduction in body segment length variance and 2° improvement in Range of Motion (ROM) angle hence enabling to more accurate measurements for upper limb exercises.
Citations
More filters
Posted Content•
TL;DR: This work introduces a Fourier basis based modeling to reconstruct the cardiovascular pulse signal at the locations containing the poor quality, that is, the locations affected by out-of-plane face movements and designs a Bayesian decision theory based HR tracking mechanism to rectify the spurious HR estimates.
Abstract: A non-invasive yet inexpensive method for heart rate (HR) monitoring is of great importance in many real-world applications including healthcare, psychology understanding, affective computing and biometrics. Face videos are currently utilized for such HR monitoring, but unfortunately this can lead to errors due to the noise introduced by facial expressions, out-of-plane movements, camera parameters (like focus change) and environmental factors. We alleviate these issues by proposing a novel face video based HR monitoring method MOMBAT, that is, MOnitoring using Modeling and BAyesian Tracking. We utilize out-of-plane face movements to define a novel quality estimation mechanism. Subsequently, we introduce a Fourier basis based modeling to reconstruct the cardiovascular pulse signal at the locations containing the poor quality, that is, the locations affected by out-of-plane face movements. Furthermore, we design a Bayesian decision theory based HR tracking mechanism to rectify the spurious HR estimates. Experimental results reveal that our proposed method, MOMBAT outperforms state-of-the-art HR monitoring methods and performs HR monitoring with an average absolute error of 1.329 beats per minute and the Pearson correlation between estimated and actual heart rate is 0.9746. Moreover, it demonstrates that HR monitoring is significantly
10 citations
Cites background from "Accurate upper body rehabilitation ..."
...[43] [9] ; and iv) observing skin damaged patients....
[...]
TL;DR: In this article, a review of state-of-the-art skeletal tracking methods using RGB-D sensors is presented, with a focus on skeletal joint kinematics analysis.
10 citations
TL;DR: The main cause of the differences between the two methods is the time required for the 3D system to acquire the data, and the involuntary body sway of human participants is more difficult to control when the time span is too long.
Abstract: Background Understanding the reliability and precision of the data obtained using three-dimensional body scanners is very important if it is intended to replace the traditional data collection methods. If the collection of anthropometric data with three-dimensional body scanners is a fast and reliable process that produces precise data at a low price, it could be used for numerous applications worldwide. Many studies have addressed data collected by white light and laser based scanners. Objective This study provides a comparative analysis between the anthropometric data collected using a Kinect body imaging system with the data collected using traditional manual methods. Moreover, a comparison is also made between the results obtained in this study and the results of previous studies of different types of body scanners. Methods The Mean Absolute Difference was calculated and all the values were compared to the maximum allowable error defined in ISO 20685. Additionally, an analysis of the significant differences between the two acquisition methods was also applied to a physical mannequin, to understand how the body movement and body stance variation in human participants impacts the results obtained. Results There are few body measurements that are close to this restricted allowable error. The results were better when the mannequin was measured. Although they were still above the ISO 20685 limit, they were much closer than the results obtained for human participants. Conclusion The main cause of the differences between the two methods is the time required for the 3D system to acquire the data. The involuntary body sway of human participants is more difficult to control when the time span is too long.
9 citations
01 Jul 2017
TL;DR: A framework for improving accuracy of Kinect skeletal tracking is proposed, that uses a set of parametric models to represent and track the human body, which indicates an improvement in accuracy of joint motion trajectories using Kinect device, rendering it more suitable for clinical assessment and rehabilitation.
Abstract: Kinect as an effective tool for clinical assessment and rehabilitation, suffers from drawbacks of lower accuracy of measuring human body kinematic data when compared to clinical gold standard motion capture devices. The accuracy of time-varying 3D locations of a fixed number of body joints obtained from Kinect skeletal tracking utility is affected by the presence of noise and precision limits of the Kinect depth sensor. In this paper, a framework for improving accuracy of Kinect skeletal tracking is proposed, that uses a set of parametric models to represent and track the human body. Each of the models represents the 3D geometric properties of a body segment connecting two adjacent joints. The temporal trajectories of the joints are recovered via particle filter-based motion tracking of each model. The proposed method was evaluated on Active Range of Motion exercises by 7 healthy subjects. The joint motion trajectories obtained using the proposed framework exhibit a greater motion smoothness (by 36%) along with reduced coefficient of variation of radius (by 34%), and lower value of root-mean-squared-error (by 53%), when compared to Kinect joint trajectories. This indicates an improvement in accuracy of joint motion trajectories using Kinect device, rendering it more suitable for clinical assessment and rehabilitation.
7 citations
Cites background or methods from "Accurate upper body rehabilitation ..."
...In contrast to these methods which utilize additional sensors, our previous work [17] improved the accuracy of Kinect skeleton data by proposing an optimization technique...
[...]
...However, the method in [17] carried out the optimization in 2D coordinate spaces of Depth and RGB, whereas the Kinect skeleton joint locations are in 3D real world coordinates....
[...]
...The optimized joint locations are derived directly in 3D coordinate space to reduce adverse effects of depth data noise present in [17]....
[...]
References
More filters
20 Jun 2011
TL;DR: This work takes an object recognition approach, designing an intermediate body parts representation that maps the difficult pose estimation problem into a simpler per-pixel classification problem, and generates confidence-scored 3D proposals of several body joints by reprojecting the classification result and finding local modes.
Abstract: We propose a new method to quickly and accurately predict 3D positions of body joints from a single depth image, using no temporal information. We take an object recognition approach, designing an intermediate body parts representation that maps the difficult pose estimation problem into a simpler per-pixel classification problem. Our large and highly varied training dataset allows the classifier to estimate body parts invariant to pose, body shape, clothing, etc. Finally we generate confidence-scored 3D proposals of several body joints by reprojecting the classification result and finding local modes. The system runs at 200 frames per second on consumer hardware. Our evaluation shows high accuracy on both synthetic and real test sets, and investigates the effect of several training parameters. We achieve state of the art accuracy in our comparison with related work and demonstrate improved generalization over exact whole-skeleton nearest neighbor matching.
3,579 citations
TL;DR: This work takes an object recognition approach, designing an intermediate body parts representation that maps the difficult pose estimation problem into a simpler per-pixel classification problem, and generates confidence-scored 3D proposals of several body joints by reprojecting the classification result and finding local modes.
Abstract: We propose a new method to quickly and accurately predict human pose---the 3D positions of body joints---from a single depth image, without depending on information from preceding frames. Our approach is strongly rooted in current object recognition strategies. By designing an intermediate representation in terms of body parts, the difficult pose estimation problem is transformed into a simpler per-pixel classification problem, for which efficient machine learning techniques exist. By using computer graphics to synthesize a very large dataset of training image pairs, one can train a classifier that estimates body part labels from test images invariant to pose, body shape, clothing, and other irrelevances. Finally, we generate confidence-scored 3D proposals of several body joints by reprojecting the classification result and finding local modes.The system runs in under 5ms on the Xbox 360. Our evaluation shows high accuracy on both synthetic and real test sets, and investigates the effect of several training parameters. We achieve state-of-the-art accuracy in our comparison with related work and demonstrate improved generalization over exact whole-skeleton nearest neighbor matching.
3,034 citations
"Accurate upper body rehabilitation ..." refers background in this paper
...in centimeters [19], and an RGB camera that captures RGB...
[...]
TL;DR: It is discussed the possibility that there is a critical level of recovery at which patients switch from a strategy employing new degrees of freedom to one in which motor recovery is produced by improving the management of degrees offreedom characteristic of healthy performance.
Abstract: A major prerequisite for successful rehabilitation therapy after stroke is the understanding of the mechanisms underlying motor deficits common to these patients. Studies have shown that in stroke patients multijoint pointing movements are characterized by decreased movement speed and increased movement variability, by increased movement segmentation and by spatial and temporal incoordination between adjacent arm joints with respect to healthy subjects. We studied how the damaged nervous system recovers or compensates for deficits in reaching, and correlated reaching deficits with the level of functional impairment. Nine right-hemiparetic subjects and nine healthy subjects participated. All subjects were right-hand dominant. Data from the affected arm of hemiparetic subjects were compared with those from the arm in healthy subjects. Seated subjects made 40 pointing movements with the right arm in a single session. Movements were made from an initial target, for which the arm was positioned alongside the trunk. Then the subject lifted the arm and pointed to the final target, located in front of the subject in the contralateral workspace. Kinematic data from the arm and trunk were recorded with a three-dimensional analysis system. Arm movements in stroke subjects were longer, more segmented, more variable and had larger movement errors. Elbow-shoulder coordination was disrupted and the range of active joint motion was decreased significantly compared with healthy subjects. Some aspects of motor performance (duration, segmentation, accuracy and coordination) were significantly correlated with the level of motor impairment. Despite the fact that stroke subjects encountered all these deficits, even subjects with the most severe motor impairment were able to transport the end-point to the target. All but one subject involved the trunk to accomplish this motor task. In others words, they recruited new degrees of freedom typically not used by healthy subjects. The use of compensatory strategies may be related to the degree of motor impairment: severely to moderately impaired subjects recruited new degrees of freedom to compensate for motor deficits while mildly impaired subjects tended to employ healthy movement patterns. We discuss the possibility that there is a critical level of recovery at which patients switch from a strategy employing new degrees of freedom to one in which motor recovery is produced by improving the management of degrees of freedom characteristic of healthy performance. Our data also suggest that stroke subjects may be able to exploit effectively the redundancy of the motor system.
862 citations
"Accurate upper body rehabilitation ..." refers background in this paper
...For example, accurate measurement of arm length is important for assessing performance of reaching task for patients with impairments in the paretic arm [16]....
[...]
TL;DR: The findings suggest that the Microsoft Kinect™ can validly assess kinematic strategies of postural control and could therefore become a useful tool for assessing posturalControl in the clinical setting.
641 citations
"Accurate upper body rehabilitation ..." refers methods in this paper
...The accuracy, validity and testretest reliability measures of Kinect have been studied for range of motion [3] [4] [5], postural control [6] [7] and gait...
[...]
TL;DR: There were consistent EMG coactivation patterns observed across all subjects (both normal and hemiparetic).
Abstract: To study abnormal spatial patterns of muscle activation in hemiparetic stroke, we compared EMG activity in paretic and contralateral elbow and shoulder muscles of 10 hemiparetic subjects during 1.5- s voluntary isometric contractions, against five to eight different loads. Isometric forces were generated in eight directions, referenced to a plane orthogonal to the long axis of the forearm, and were recorded by a three degrees of freedom load cell, mounted at the wrist. Surface and intramuscular EMGs of six elbow and six shoulder muscles were recorded from both impaired and contralateral upper extremities of each subject. The spatial characteristics of EMG activation of individual muscles were summarized using two measures. The first, called the ‘net resultant EMG vector’ is a new measure which calculated the vector sum of EMG magnitudes for each of the eight directions, and the second, index of EMG focus, is a measure of the range of EMG activation recorded for each load level. Use of these measures permitted us to describe spatial EMG characteristics quantitatively, which has not been done previously. We observed consistent and statistically significant shifts in the resultant EMG vector directions in the impaired limb, especially in shoulder and other proximal muscles. Significant increases in the angular range of EMG activity were also identified and were most evident at the elbow. Correlation analysis techniques were used to assess the degree of coactivation of different muscle pairs. There were consistent EMG coactivation patterns observed across all subjects (both normal and hemiparetic). However, in spastic-paretic limbs, additional novel coactivational relationships were also recorded, especially between elbow flexors/shoulder abductors and elbow extensors/shoulder adductors. These novel coactivation patterns represent a reduction in the number of possible muscle combinations, or in the number of possible ‘synergies’ in the paretic limb of the stroke subject. This reduction in number of ‘synergies’ could result from a loss of descending command options; from an increased reliance on residual, descending brainstem pathways (such as the reticulospinal and vestibulospinal projections); from changes in spinal interneuronal excitability; or from a combination of several of these factors. The relative merits of these hypotheses are addressed.
616 citations
"Accurate upper body rehabilitation ..." refers background in this paper
...Several studies suggest that abnormal coupling of shoulder adductors with elbow extensors and shoulder abductors with elbow flexors often leads to some stereotypical movement characteristics exhibited by severe stroke patients [2]....
[...]