Claims:1. A processor implemented method (200) comprising:
receiving, from a 3-Dimensional (3D) motion sensor, skeleton data for a plurality of joints in static and dynamic postures over a plurality of ranges of motion, the skeleton data comprising measured 3D joint coordinates (202); and
tracking the plurality of joints by a constrained linear dynamic filtering technique based on a state space model, wherein the measured 3D joint coordinates of each pair of physically connected joints from the plurality of joints, are filtered simultaneously, with a constraint that bone length connecting each pair of the physically connected joints remains constant, to obtain optimized 3D joint coordinates, the bone length being the Euclidean distance between each pair of the physically connected joints (204).

2. The processor implemented method of claim 1 further comprising tracking dynamic trajectory of the plurality of joints based on the optimized 3D joint coordinates (206).

3. The processor implemented method of claim 1, wherein the constrained linear dynamic filtering is a constrained Kalman filtering technique.

4. The processor implemented method of claim 1, wherein the step of tracking comprises:
forming a state vector and measurement vector constituting the state space model, based on the measured 3D joint coordinates of each pair of the physically connected joints for each instant of time;
estimating 3D joint coordinates of each pair of the physically connected joints for a next state based on a previous state of the 3D joint coordinates by minimizing a mean square error between the estimated 3D joint coordinates and the measured 3D joint coordinates based on a pre-defined gain to obtain updated 3D joint coordinates; and
applying the constraint that the bone length connecting each pair of the physically connected joints remains constant to generate the optimized 3D joint coordinates.

5. A system (100) comprising:
one or more data storage devices (102) operatively coupled to one or more hardware processors (104) and configured to store instructions configured for execution by the one or more hardware processors to:
receive, from a 3-Dimensional (3D) motion sensor, skeleton data for a plurality of joints in static and dynamic postures over a plurality of ranges of motion, the skeleton data comprising 3D joint coordinates; and
track the plurality of joints by a constrained linear dynamic filtering technique based on a state space model, wherein the measured 3D joint coordinates of each pair of physically connected joints from the plurality of joints, are filtered simultaneously, with a constraint that bone length connecting each pair of the physically connected joints remains constant, to obtain optimized 3D joint coordinates, the bone length being the Euclidean distance between each pair of the physically connected joints.

6. The system of claim 5, wherein the one or more hardware processors are further configured to track dynamic trajectory of the plurality of joints based on the optimized 3D joint coordinates.

7. The system of claim 5, wherein the constrained linear dynamic filtering is a constrained Kalman filtering technique.

8. The system of claim 5, wherein the one or more hardware processors are further configured to track the plurality of joints by:
forming a state vector and measurement vector constituting the state space model, based on the measured 3D joint coordinates of each pair of the physically connected joints for each instant of time;
estimating 3D joint coordinates of each pair of the physically connected joints for a next state based on a previous state of the 3D joint coordinates by minimizing a mean square error between the estimated 3D joint coordinates and the measured 3D joint coordinates based on a pre-defined gain to obtain updated 3D joint coordinates; and
applying the constraint that the bone length connecting each pair of the physically connected joints remains constant to generate the optimized 3D joint coordinates.
, Description:FORM 2

THE PATENTS ACT, 1970
(39 of 1970)
&
THE PATENT RULES, 2003

COMPLETE SPECIFICATION
(See Section 10 and Rule 13)

Title of invention:
CONSTRAINED LINEAR DYNAMIC FILTERING FOR
ANTHROPOMETRIC MEASUREMENTS

Applicant:
Tata Consultancy Services Limited
A company Incorporated in India under the Companies Act, 1956
Having address:
Nirmal Building, 9th floor,
Nariman point, Mumbai 400021,
Maharashtra, India

The following specification particularly describes the embodiments and the manner in which it is to be performed.

TECHNICAL FIELD
[001] The embodiments herein generally relate to joint motion analysis, and more particularly to systems and methods for constrained linear dynamic filtering for anthropometric measurements.

BACKGROUND
[002] Joint motion analysis finds application in the gaming industry, physical medicine and rehabilitation domain. Anthropometric measures are critical to joint motion analysis and form an important aspect of health monitoring of patients suffering from neurological disorders, post-stroke patients, and elderly subjects. Marker based joint movement data capture using clinical gold standard devices such as Vicon™, are extremely expensive and hence may not be viable for prolonged rehabilitation therapy which may be absolutely essential for stroke patients. Kinect™ is a marker-less motion capture device that is being used for its low cost and portability when compared to its expensive counterpart. The accuracy of Kinect™ implemented by Xbox 360™ as well as Xbox one™ as an alternative to marker-based systems have indicated a disparity in accuracy of the human body joint information provided by Kinect™ version1(Xbox 360™ Kinect sensor) and Kinect™ version2 (Xbox One™ Kinect sensor) skeletal data. High frequency fluctuations in joint locations and error in tracking during dynamic conditions make joint information unreliable for clinical or other high precision applications.

SUMMARY
[003] Embodiments of the present disclosure present technological improvements as solutions to one or more of the above-mentioned technical problems recognized by the inventors in conventional systems.
[004] In an aspect, there is provided a processor implemented method comprising: receiving, from a 3-Dimensional (3D) motion sensor, skeleton data for a plurality of joints in static and dynamic postures over a plurality of ranges of motion, the skeleton data comprising measured 3D joint coordinates; and tracking the plurality of joints by a constrained linear dynamic filtering technique based on a state space model, wherein the measured 3D joint coordinates of each pair of physically connected joints from the plurality of joints, are filtered simultaneously, with a constraint that bone length connecting each pair of the physically connected joints remains constant, to obtain optimized 3D joint coordinates, the bone length being the Euclidean distance between each pair of the physically connected joints.
[005] In another aspect, there is provided a system comprising: one or more data storage devices operatively coupled to the one or more processors and configured to store instructions configured for execution by the one or more processors to: receive, from a 3-Dimensional (3D) motion sensor, skeleton data for a plurality of joints in static and dynamic postures over a plurality of ranges of motion, the skeleton data comprising measured 3D joint coordinates; and track the plurality of joints by a constrained linear dynamic filtering technique based on a state space model, wherein the measured 3D joint coordinates of each pair of physically connected joints from the plurality of joints, are filtered simultaneously, with a constraint that bone length connecting each pair of the physically connected joints remains constant, to obtain optimized 3D joint coordinates, the bone length being the Euclidean distance between each pair of the physically connected joints.
[006] In yet another aspect, there is provided a computer program product comprising a non-transitory computer readable medium having a computer readable program embodied therein, wherein the computer readable program, when executed on a computing device, causes the computing device to: receive, from a 3-Dimensional (3D) motion sensor, skeleton data for a plurality of joints in static and dynamic postures over a plurality of ranges of motion, the skeleton data comprising measured 3D joint coordinates; and track the plurality of joints by a constrained linear dynamic filtering technique based on a state space model, wherein the measured 3D joint coordinates of each pair of physically connected joints from the plurality of joints, are filtered simultaneously, with a constraint that bone length connecting each pair of the physically connected joints remains constant, to obtain optimized 3D joint coordinates, the bone length being the Euclidean distance between each pair of the physically connected joints.
[007] In an embodiment of the present disclosure, the one or more hardware processors are further configured to track dynamic trajectory of the plurality of joints based on the optimized 3D joint coordinates.
[008] In an embodiment of the present disclosure, the one or more hardware processors are further configured to track the plurality of joints by: forming a state vector and measurement vector constituting the state space model, based on the measured 3D joint coordinates of each pair of the physically connected joints for each instant of time; estimating 3D joint coordinates of each pair of the physically connected joints for a next state based on a previous state of the 3D joint coordinates by minimizing a mean square error between the estimated 3D joint coordinates and the measured 3D joint coordinates based on a pre-defined gain to obtain updated 3D joint coordinates; and applying the constraint that the bone length connecting each pair of the physically connected joints remains constant to generate the optimized 3D joint coordinates.
[009] In an embodiment of the present disclosure, the constrained linear dynamic filtering is a constrained Kalman filtering technique.
[010] It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the embodiments of the present disclosure, as claimed.

BRIEF DESCRIPTION OF THE DRAWINGS
[011] The embodiments herein will be better understood from the following detailed description with reference to the drawings, in which:
[012] FIG.1 illustrates an exemplary block diagram of a system for constrained linear dynamic filtering for anthropometric measurements, in accordance with an embodiment of the present disclosure;
[013] FIG.2 is an exemplary flow diagram illustrating a computer implemented method for constrained linear dynamic filtering for anthropometric measurements, in accordance with an embodiment of the present disclosure;
[014] FIG.3 illustrates a schematic representation of joints tracked by Kinect™ version2;
[015] FIG.4 illustrates a comparative graphical representation of (i) tracked right elbow joint coordinates obtained from Kinect™ version2 for shoulder abduction/adduction, (ii) joint coordinates obtained by unconstrained Kalman filter and (iii) joint coordinates obtained by constrained Kalman filter in accordance with an embodiment of the present disclosure; and
[016] FIG.5 illustrates a comparative graphical representation of (i) unfiltered arm length obtained from Kinect™ version2 for shoulder abduction/adduction, (ii) arm length obtained by unconstrained Kalman filter and (iii) arm length obtained by constrained Kalman filter in accordance with an embodiment of the present disclosure.
[017] It should be appreciated by those skilled in the art that any block diagram herein represent conceptual views of illustrative systems embodying the principles of the present subject matter. Similarly, it will be appreciated that any flow charts, flow diagrams, state transition diagrams, pseudo code, and the like represent various processes which may be substantially represented in computer readable medium and so executed by a computing device or processor, whether or not such computing device or processor is explicitly shown.
DETAILED DESCRIPTION
[018] The embodiments herein and the various features and advantageous details thereof are explained more fully with reference to the non-limiting embodiments that are illustrated in the accompanying drawings and detailed in the following description. The examples used herein are intended merely to facilitate an understanding of ways in which the embodiments herein may be practiced and to further enable those skilled in the art to practice the embodiments herein. Accordingly, the examples should not be construed as limiting the scope of the embodiments herein.
[019] The words "comprising," "having," "containing," and "including," and other forms thereof, are intended to be equivalent in meaning and be open ended in that an item or items following any one of these words is not meant to be an exhaustive listing of such item or items, or meant to be limited to only the listed item or items.
[020] It must also be noted that as used herein and in the appended claims, the singular forms "a," "an," and "the" include plural references unless the context clearly dictates otherwise. Although any systems and methods similar or equivalent to those described herein can be used in the practice or testing of embodiments of the present disclosure, the preferred, systems and methods are now described.
[021] Some embodiments of this disclosure, illustrating all its features, will now be discussed in detail. The disclosed embodiments are merely exemplary of the disclosure, which may be embodied in various forms.
[022] Before setting forth the detailed explanation, it is noted that all of the discussion below, regardless of the particular implementation being described, is exemplary in nature, rather than limiting.
[023] Marker-less motion sensing devices such as Microsoft Kinetic™ are very popular due to its affordability, portability and unobtrusiveness. However, skeleton data obtained from such devices are usually noisy which affects accuracy of estimation of 3-Dimensional (3D) joint locations. The noise profile varies for both stationary and dynamic postures and it affects anthropometric measurements of the body segments connecting any two joints. This may not be particularly suitable for applications in the health and rehabilitation industry. The present disclosure provides a constrained linear dynamic filter based approach that not only tracks the joints accurately but also reduces the variation in bone lengths observed in conventional methods.
[024] In the context of the present disclosure, the expression “ranges of motion” refers to any kind of movement of a body during dynamic postures involving joints in the body such as ROM (Range of Motion) exercises or exercise such as walking or sit to stand, and the like. Although the present disclosure may be explained with reference to skeleton data received particularly from Microsoft Kinetic™, it may be understood that skeleton data may be received from any 3D motion sensing device. The expressions Kinect™, Kinect™ version1, Kinect™ version2 may be used interchangeably in the description hereunder to represent the motion sensing device used in an exemplary embodiment; Kinect™ version1 and Kinect™ version2 represent two versions of the Microsoft™ product.
[025] Referring now to the drawings, and more particularly to FIGS. 1 through 5, where similar reference characters denote corresponding features consistently throughout the figures, there are shown preferred embodiments and these embodiments are described in the context of the following exemplary system and method.
[026] FIG.1 illustrates an exemplary block diagram of a system 100 for constrained linear dynamic filtering for anthropometric measurements, in accordance with an embodiment of the present disclosure. In an embodiment, the system 100 includes one or more processors 104, communication interface device(s) or input/output (I/O) interface(s) 106, and one or more data storage devices or memory 102 operatively coupled to the one or more processors 104. The one or more processors 104 that are hardware processors can be implemented as one or more microprocessors, microcomputers, microcontrollers, digital signal processors, central processing units, state machines, graphics controllers, logic circuitries, and/or any devices that manipulate signals based on operational instructions. Among other capabilities, the processor(s) are configured to fetch and execute computer-readable instructions stored in the memory. In an embodiment, the system 100 can be implemented in a variety of computing systems, such as laptop computers, notebooks, hand-held devices, workstations, mainframe computers, servers, a network cloud and the like.
[027] The I/O interface device(s) 106 can include a variety of software and hardware interfaces, for example, a web interface, a graphical user interface, and the like and can facilitate multiple communications within a wide variety of networks N/W and protocol types, including wired networks, for example, LAN, cable, etc., and wireless networks, such as WLAN, cellular, or satellite. In an embodiment, the I/O interface device(s) can include one or more ports for connecting a number of devices to one another or to another server.
[028] The memory 102 may include any computer-readable medium known in the art including, for example, volatile memory, such as static random access memory (SRAM) and dynamic random access memory (DRAM), and/or non-volatile memory, such as read only memory (ROM), erasable programmable ROM, flash memories, hard disks, optical disks, and magnetic tapes. In an embodiment, one or more modules (not shown) of the system 100 can be stored in the memory 102.
[029] FIG.2 illustrates an exemplary flow diagram illustrating a computer implemented method 200 for constrained linear dynamic filtering for anthropometric measurements, in accordance with an embodiment of the present disclosure. In an embodiment, the system 100 comprises one or more data storage devices or memory 102 operatively coupled to the one or more processors 104 and is configured to store instructions configured for execution of steps of the method 200 by the one or more processors 104.
[030] In an embodiment, at step 202, the one or more processors 104 are configured to receive skeleton data for a plurality of joints in static and dynamic postures over a plurality of ranges of motion. The skeleton data comprises measured 3-Dimensional (3D) joint coordinates. Kinect™ provides noisy 3D joint coordinates (ai, bi, ci) for i = 1, 2, ..N, where N is the total number of joints tracked by the Kinect™ and a, b, c are three coordinate axes in 3D space. In an exemplary embodiment, for Kinect™ version1, N = 20 and for Kinect™ version2, N = 25
[031] In an embodiment at step 204, the one or more processors 104 are configured to track the plurality of joints by a constrained linear dynamic filtering technique based on a state space model, wherein the measured 3D joint coordinates of each pair of physically connected joints are filtered simultaneously. In an embodiment, the constrained linear dynamic filtering technique may be a Kalman filter, a Weiner filter technique, and the like that may be constrained as explained herein below.
[032] In accordance with the present disclosure, the filtering is performed with a constraint that bone length connecting each pair of the physically connected joints remains constant, to obtain optimized 3D joint coordinates. The bone length under consideration is the Euclidean distance between each pair of physically connected joints. FIG.3 illustrates a schematic representation of joints tracked by Kinect™ version2. In accordance with the present disclosure, firstly, a state vector is formed as x = {ai, aj, bi, bj, ci, cj]T, where i, j are two physically connected joint numbers as illustrated in FIG.3. For instance, elbow joints (joints 5 and 9) and wrist joints (joints 6 and 10) can form a state vector as they are connected through forearm bone whereas elbow and hip joint (joint 0) cant form a state vector as they are not physically connected.
[033] In accordance with an embodiment, for applying Kalman filtering, motion of the limbs in the skeleton is modeled as a linear dynamic system where the next state at time instance k+1 is expressed in terms of previous state at kth instance and is expressed as
xk+1 = Fxk + Qk ? (1)
yk = Hxk + Rk ? (2)
where xk and yk are the state vector and measurement vector constituting the state space model respectively at time instant k, having dimension [6X1]. Qk and Rk are the process noise and measurement noise respectively whereas F and H are the state transition matrix and state transformation matrix respectively of dimension [6X6].
[034] Kalman filter estimates (unconstrained estimation) from xk having the knowledge of the measurement vector yk = {y0, y1,..yk} in two steps namely prediction and update. The standard Kalman filtering prediction step can be written as
k = F k-1 ? (3)
k = F Pk-1 FT + Qk ? (4)
where having dimension [6X6] is a covariance matrix associated with the prediction k for an unknown true state xk and can be expressed as
= E [(xk - k) (xk - k) T] ? (5)
The updated state based on the measurement is expressed as
Kk = k HT (H k HT + R)-1 ? (6)
= +Kk (yk-H k) ? (7)
Pk= (I - Kk H) k ? (8)
Here K is the Kalman gain matrix of dimension [6X6] and Q, R are process noise and measurement noise. Here Pk having dimension [6X6] is the covariance associate with the updated state vector k. Kalman filter minimizes the mean square error between estimated k and true xk providing smoother updated 3D joint coordinates. However in the above formulation the variation of bone length between two physically connected joints is not taken into consideration. The bone length l between two joints i and j can be expressed as Euclidean distance between the 3D coordinates of those two joints and is expressed as

?(9)
Thus in order to keep l constant over time, the non-linear constraint at kth instance is defined in matrix form as
f (xk ) = xk A - l2 = 0 ? (10)

where A constructed using equation (9) as

? (11)

In order to keep the bone length fixed, the unconstrained state estimation k is projected on the constrained surface as given in equation (10). Hence an objective function for this optimization problem is formulated as:
? (12)

where W is a positive symmetric matrix and is the constrained estimation of xk. For solving equation (12) Lagrangian Multiplier technique is applied. The Lagrangian is written as:
? f ( ) ? (13)
where ? is the Lagrangian multiplier. In order to find and ?, the first order derivatives of G with respect to xk and ? respectively are computed and are set to zero. i.e.

? (14)
The constrained solution can be obtained from equation (14) and is written as
? (15)
In equation (15), W is taken as P-1 as it provides smaller covariance of compared to the covariance of unconstrained estimation . In equation (15) the only unknown is ?. Hence is replaced in equation (10) to obtain ?. In order to calculate ?, equation (15) is written as
? (16)
Where L is a [6 X 6] matrix obtained from Cholesky decomposition of W. Since W is a positive definite matrix, hence it is written as
W = L* L ? (17)
Where L* is a complex conjugate of L. As we are dealing with the real matrix so the complex conjugate can be replaced by the transpose of the matrix. Again in equation (16), M is a normal matrix of dimension [6X6] that diagonalizes the matrix L-1TAL-1. Hence J is the diagonal matrix of dimension [6X6], given by
J = ML-1T AL-1 MT ? (18)
Further, equation (16) may be re-written by applying matrix inversion identity as given in equation (19)
? (19)
Replacing equation (19) in equation (10) and expressing equation (10) in scalar form, q(?) is expressed as
? (20)
where Now the solution for ? can be obtained iteratively by Newton’s method starting from ? = 0. Mathematically it is written as
? (21)
The value of ? obtained from equation (21) is put in equation (19) to obtain the constrained state estimation for joints i and j. This process is repeated for all joints to obtain optimized 3D joint data keeping Euclidean distance between two physically connected joints constant.
[035] In an embodiment, the method 200 further comprises tracking dynamic trajectory of the plurality of joints based on the optimized 3D joint coordinates at step 206.

EXPERIMENTAL SETUP AND RESULTS
[036] Experimental setup: Database creation:
Twenty five subjects (age: 24-60 years, weight: 50kg-100kg and height: 1.42m-1.96m) with no symptoms of neuro-physiological disorder, had been chosen for the study. Both Kinect™ version1 and Kinect™ version2 were used for capturing skeleton data (3D coordinates of 20 and 25 joints for Kinect™ version1 and Kinect™ version2 respectively) in two successive sessions as simultaneous data capture is not possible due to heavy IR interference between two devices placed side by side. Thus subjects performed active Range Of Motion (ROM) exercises - shoulder abduction/adduction and flexion/extension in succession in front of the Kinect™ placed at a distance of 3m. In the beginning of the exercise, the subjects were told to stand in stationary posture for 30 seconds and then perform the exercise. Dataset comprises of skeleton joint coordinates for both static and dynamic postures as noise characteristics for static and dynamic postures might be different. The intent of the experiment was to verify the method of the present disclosure for both static and dynamic postures using data set obtained from different versions of Kinect™.
[037] Results and discussion
Performance of the method of the present disclosure needed to be evaluated for joint tracking as well as for assessing improvements in bone length’s statistics during both static and dynamic postures obtained by filtering of corresponding static and dynamic joint coordinates. As an example, for right shoulder abduction/adduction exercise, joints namely shoulder, elbow, wrist and hand are the dynamic joints and corresponding joints in the left arm are considered as static joints. Similarly as shown in FIG.3, bones connecting joints 8 and 9 (i.e. arm), joints 9 and 10 (i.e. forearm) and joints 10 and 11 (i.e. palm) are corresponding to dynamic body segments of right hand. For left hand corresponding bones are considered as static joints. FIG.4 illustrates a comparative graphical representation of (i) tracked right elbow joint coordinates obtained from Kinect™ version2 for shoulder abduction/adduction, (ii) joint coordinates obtained by unconstrained Kalman filter and (iii) joint coordinates obtained by constrained Kalman filter in accordance with an embodiment of the present disclosure. FIG.4 shows that both the unconstrained and constrained Kalman filter of the present disclosure are equally capable of tracking the joint location obtained from Kinect™ version2 but the unconstrained Kalman filter does not preserve the bone length (FIG.5). Moreover, zoomed portions of b coordinates in FIG.4 clearly depict that the method of the present disclosure does not abruptly modify the joint coordinates to satisfy the constraint, rather a constrained optimization is done based on the accurate joint position tracking. The observation also holds true for a and c coordinates. FIG.5 illustrates a comparative graphical representation of (i) unfiltered arm length obtained from Kinect™ version2 for shoulder abduction/adduction, (ii) arm length obtained by unconstrained Kalman filter and (iii) arm length obtained by constrained Kalman filter in accordance with an embodiment of the present disclosure. FIG.5 proves the superiority of the method of the present disclosure (constrained Kalman filter here) in reducing the variation in bone length as compared to the same calculated from unfiltered coordinates (obtained from Kinect™ version2) and filtered coordinates output from the unconstrained Kalman filter. It is clearly seen from FIG.5 that in both stationary (U1 - U2 and U3 - U4) and dynamic (U2 - U3 and U4 - U5) phases, the constrained Kalman filter works better than the unconstrained version, however in dynamic phase this improvement is much more prevalent. Moreover, between two dynamic regions U2 - U3 and U4 - U5, the second static region U3 - U4 is of different posture than the first static one U1 - U2. In these two different static postures (U1 - U2 and U3 - U4), the bone length varies for the unfiltered data and the unconstrained Kalman filter is not able to correct the same. However, the constrained Kalman filter preserves the bone length irrespective of the static posture.
[038] The performance evaluation for both static and dynamic body postures have been reported based on the standard deviation (STD) of bone length between every physically connected joints. Ideally the variation in bone length is expected to be zero, so less the STD, better is the performance. Table I shows average (mean) of STD of bone lengths computed for all the subjects while doing right shoulder abduction/adduction.
Table I: Variation (STD) in arm, forearm and palm lengths in meters for Kinect™ version2 (for shoulder abduction/adduction)
Segments Unfiltered Unconstrained Kalman Constrained Kalman
Right arm 0.0163 0.0169 0.0004
Right forearm 0.0133 0.0130 0.0005
Right palm 0.0116 0.0093 0.0028
Left arm 0.0039 0.0032 0.0001

The joint coordinates obtained from Kinect™ version2 have been filtered using unconstrained and constrained Kalman filter and is presented in 3rd and 4th columns of Table I. It is clear from Table I that according to method of the present disclosure, bone length variation over time is the least. It can also be observed that, as left arm is a static bone, so the corrected bone length estimated by the method of the present disclosure is almost equal to zero. On the other hand, for dynamic bones, there is a substantial improvement for all three body segments. Thus it is clear that the constrained Kalman filter is capable of tracking both static and dynamic joints by preserving the bone length over time.
[039] Similar results are generated for Kinect™ version1 data and are reported in Table II.
Table II: Variation (STD) in arm, forearm and palm lengths in meters for Kinect™ version1 (for shoulder abduction /adduction)
Segments Unfiltered Unconstrained Kalman Constrained Kalman
Right arm 0.0165 0.0231 0.0008
Right forearm 0.0393 0.0236 0.0014
Right palm 0.0949 0.0757 0.0228
Left arm 0.0058 0.0053 0.0002
However, as the right palm is connecting two most noisy joints (compared to right arm and forearm) the absolute STD for the method of the present disclosure (Table II, 4th column - 4th row) still remains high.
[040] Improvements were seen for 3 dynamic (i.e. right arm, forearm and palm) and one static bone (i.e. left arm) lengths in Table I and Table II but the method of the present disclosure performs well for all other bones and is capable of tracking the corresponding joints properly for all exercises. Finally the evaluation as done over the entire dataset using
= mean ( )
where algo = {unconstrained, constrained} Kalman filter. More the value of (? ) better is the result. For Kinect™ version2, the method of the present disclosure reduces STD of all bone lengths computed from all the joints over all the subjects by = 92% and same for Kinect™ version1 is 94%. On the other hand is 10% and 18% for Kinect™ version2 and Kinect™ version1 respectively considering unconstrained Kalman filter based estimation. Thus the method of the present disclosure outperforms unconstrained Kalman filtering approach for smoothing of joint coordinates by tracking joint positions accurately keeping bone length constant.
[041] The present disclosure thus provides systems and methods to improve the anthropometric measurements of skeleton data obtained from motion sensing devices such as Kinect™ which may be used in rehabilitation, gaming and such other applications that are dependent on joint motion analysis. The method of the present disclosure is capable of tracking the dynamic trajectory of joints by keeping the bone length (distance between two physically connected joints) constant over time. The method of the present disclosure has been evaluated for both static and dynamic body posture data collected using both Kinect™ version1 and Kinect™ version2. The results provided herein above show that it provides accurate tracking of dynamic joints with significant minimization of bone length variations. Although in the present disclosure, the constraint has been incorporated into a constant value Kalman filter model, it may be extended to other variants of Kalman filter.
[042] The written description describes the subject matter herein to enable any person skilled in the art to make and use the embodiments of the present disclosure. The scope of the subject matter embodiments defined here may include other modifications that occur to those skilled in the art. Such other modifications are intended to be within the scope if they have similar elements that do not differ from the literal language of the claims or if they include equivalent elements with insubstantial differences from the literal language.
[043] The scope of the subject matter embodiments defined here may include other modifications that occur to those skilled in the art. Such other modifications are intended to be within the scope if they have similar elements that do not differ from the literal language of the claims or if they include equivalent elements with insubstantial differences from the literal language.
[044] It is, however to be understood that the scope of the protection is extended to such a program and in addition to a computer-readable means having a message therein; such computer-readable storage means contain program-code means for implementation of one or more steps of the method, when the program runs on a server or mobile device or any suitable programmable device. The hardware device can be any kind of device which can be programmed including e.g. any kind of computer like a server or a personal computer, or the like, or any combination thereof. The device may also include means which could be e.g. hardware means like e.g. an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or a combination of hardware and software means, e.g. an ASIC and an FPGA, or at least one microprocessor and at least one memory with software modules located therein. Thus, the means can include both hardware means and software means. The method embodiments described herein could be implemented in hardware and software. The device may also include software means. Alternatively, the embodiments of the present disclosure may be implemented on different hardware devices, e.g. using a plurality of CPUs.
[045] The embodiments herein can comprise hardware and software elements. The embodiments that are implemented in software include but are not limited to, firmware, resident software, microcode, etc. The functions performed by various modules comprising the system of the present disclosure and described herein may be implemented in other modules or combinations of other modules. For the purposes of this description, a computer-usable or computer readable medium can be any apparatus that can comprise, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device. The various modules described herein may be implemented as software and/or hardware modules and may be stored in any type of non-transitory computer readable medium or other storage device. Some non-limiting examples of non-transitory computer-readable media include CDs, DVDs, BLU-RAY, flash memory, and hard disk drives.
[046] Further, although process steps, method steps, techniques or the like may be described in a sequential order, such processes, methods and techniques may be configured to work in alternate orders. In other words, any sequence or order of steps that may be described does not necessarily indicate a requirement that the steps be performed in that order. The steps of processes described herein may be performed in any order practical. Further, some steps may be performed simultaneously.
[047] The preceding description has been presented with reference to various embodiments. Persons having ordinary skill in the art and technology to which this application pertains will appreciate that alterations and changes in the described structures and methods of operation can be practiced without meaningfully departing from the principle, spirit and scope.