3D Motion Analysis - Science topic

To compare the path of two spherical pendulums in 3D space, you can follow these steps:

First, define the equations of motion for each pendulum. These equations will describe the position and orientation of each pendulum as a function of time.

Solve these equations of motion numerically using a computer program or mathematical software such as MATLAB or Python. This will give you a set of data points that describe the trajectory of each pendulum over time.

Plot the trajectory of each pendulum in 3D space. You can use a 3D plotting software such as MATLAB's "plot3" or Python's "matplotlib" library to create these plots.

Compare the trajectories of the two pendulums visually by overlaying the plots. You can adjust the colors and styles of the plots to make them easier to distinguish.

Analyze the trajectories of the two pendulums by comparing their shapes, amplitudes, frequencies, and other characteristics. You can also use statistical techniques such as correlation analysis to quantify the similarity between the two trajectories.

Draw conclusions based on your analysis. Depending on your research question, you may be interested in identifying similarities or differences between the two pendulums, or in understanding the factors that influence their motion.

Hello Phi,

If it's just pairwise relationships you're interested in, then gamma or spearman rho could suffice. If, on the other hand, you were interested in what ensemble of the "emotion" variables accounts for differences on the reported intent to purchase, then ordinal regression might be suitable.

Good luck with your work.

Hi Alireza,

First, what is the type of motion capture system do you have? Most of them actually have their own synchronization box, which serve as trigger in or trigger out for other external device including IMU. However, if your IMU sensor is a standalone sensor without any base to connected, then you can try an option to synchronize it by using some movement that cause spike on IMU signal like jumping.

Hope this helps you

Dear Mohammed,

maybe it would help to make your question a bit more specific. The plots seems to refer to solution using conformal mapping.

Please mention the reference (in English) from which you took the figure and the short part of the text defining the amplitude ratio.

For some classical references you can take a look at these:

Lewis, F.M., 1929. The inertia of the water surrounding a vibrating ship, in: 37th Meeting of The Society of Naval Architects and Marine Engineers, SNAME.

Ursell, B.F., 1949. On the heaving motion of a circular cylinder. Q. J. Mech. Appl. Math. II, 218–231.

Frank, W., 1967. Oscillation of cylinders in or below the free surface of deep fluids. Naval Ship Research and Development Center, Washington

Related to your query, you may use some easy Matlab., e.g., https://www.mathworks.com/help/fusion/ref/imufilter-system-object.html

Kinovea

Hello;

I agree that is the most used in that case.ht the software : EthoVision XT

Thank you for the answer Dr. Jokstad

Best,

Jose Heredia-Jimenez

I am using Vicon system and trying to answer you according to Vicon.

Your concern is correct. The static frame will be used to zero the offset angle and regard it as the anatomical neutral position. If you do the standard procedure for normal subject, it will be wrong for abnormal offset angles.

If possible, you may do the calculation by "Body Builder", a software for Vicon data modelling.

If not, you may import the trails after intervention into the old data session. You trick the system they are data in the same session. Then, the angle change will be the difference between the old and new static frame.

Hope it helps

Thank's for your answers Fokko and Galya. As I understand you correctly, for your mentioned applications an average cycle or even for looking only onto the heart rate a simple ECG would be enough. I still wonder, if real time resolved 3D images over several respiratory or cardiac cycles are relevant for diagnosis. Or would that open new diagnostic opportunities?

I suppose that your IVs are continuos variables.

Thus you can use determinant analysis: this will predict the appartenance of participants to one of the three groups based on the scores obtained on the 4 IVs.

regardless of test method, is there possibility of having this kind of clinical scenario??

From what you have written, I assume that these points you are referring to are 3D coordinates of markers placed on the subject during weightlifting task. Also I am assuming that these points are labelled such that you have a continuous array of 3D coordinates for each marker.

There will be some error in the detected positions due to noise. There are generally two ways how to filter out this noise, either by using Butterworth filter (en.wikipedia.org/wiki/Butterworth_filter) with typically 10 Hz cutoff frequency. You can adjust the cutoff frequency according to the task itself - slower movements = higher cutoff frequency.

The second method to filter the noise is to use Savitzky-Golay method (en.wikipedia.org/wiki/Savitzky–Golay_filter). Here you will have to adjust the order and the frame length. These values are highly dependent on your data and you will just have to test few values to see which give the best results.

Once you filter the data, you can then proceed with obtaining the velocity the usual way, i.e. "velocity = diff(position)./diff(time)" as you wrote.

Hei,

I am so sorry for not answering you sooner but I did not have time yet.... ☹️ It is good to hear that you managed to solve the problem! 😀

Have a nice day!

Annamaria

thanks David and Alasdair

Hi All, 

Like what Senthil is saying, there is no real gold standard but the closest to that would be motion capture systems such as Vicon or optotrak.  To compare your method you will need to capture the same motion using one gold standard and your technique. In addition you can run these motions multiple times to test repeatability. Then you can get the RMS error or standard deviation as measures of how tour technique measures with these gold standards. Your technique could match the accuracy of gold standards but offer measurement outside structured environments. 

Hi

See the given paper, you can get your answer

Vasanth,

The marker placement strategy can certainly dictate how you determine the underlying skeletal kinematics.  There are, for example, the point cluster method by Andriacchi et al and minimum marker (less than 3 per body segment) method we developed which would use mathematical optimization to enforce the rigidity of underlying linkage.  Hope this helps.

Hope some of this can help. Especially work of Mr.Ertan.

You can also look for unpublished doctoral thesis "The analysis of shooting dynamics in archery" by Tinazci C.  and  "Total Archery" book (Samick Sports) by Lee K. and Bondt R. 

I hope You will find some of this information helpful.

Kind regards,

Jan Homolak

Hello Peter,

I would say that it depends on the level of accuracy you want to get and the general purpose of computing these variables (for discriminating different phases of a particular movement, f.ex.). If these are simple kinematic quantities (angular motion of joint angles or limb accelerations...) ,  I think you can either use the upper limb models you mentioned. The variations in the limb/joints during movement execution would follow a pattern close to real, I assume. Otherwise, I would recommend to create your own model (for throwing) by yourself in Nexus.

I would be curious to see the distortions between such a model and the ISB standards.

Halim

Sorry - I inluded the wrong link in my previous post. The link should be:

The link I mistakenly put might also be useful to you however, since it deals with the related problem of tracking objects whose trajectories are not straight (i.e. that involve turns, etc.).

In Abaqus  solver, load (pressure) can be applied as a function of time and space. For defining space variation, you need to define it as a function. However there are various ways in which one can define pressure/time relation. you can add any amplitude (Pressure vs time ) data as a CAE input. amplitude can be added in number of ways like tabular, periodic etc.

A procedural animation is any animation controlled by programmer goal, not real-life one. Examples: procedural generation of plants, particle systems, movement along motion graph, etc. Often such applications are designed just to provide nice look and widely use random elements. But still in the simulation Physics and Dynamics can be used.

If planning of the next group of steps in your case of locomotion depends on game events and interaction with the scene objects, this is not “procedural-based” animation. If the actor or animals walk through the scene in random direction this is a kind of procedure generated animation. In any case algorithm of locomotion simulation still is based on Dynamics or Inverse Kinematics.

I am not sure whether you want to know about software or hardware configuration settings. There is a Quick Start Guide that comes along with the Users Reference manual that describes the software settings and how to aim and focus the Osprey cameras. Appendix A in the Reference Manual describes how to connect the cameras to a computer.

hello Nikolaos

I used Kinovea (freeware) to retrieve the target coordinates in 2D on both camera.

Then I concatenate two 2D data with a DLT (direct linear transformation) to obtain 3D with matlab, labview or openCV (freeware) script

Abdel- Aziz Y. I, Karara H.M., 1971. Direct Linear Transformation from Comparator Coordinates into ObjectSpace Coordinates.ASP Symposium on Close- Range Photogrammtery, Falls Church, VA, pp.1-18.

enjoy

Patrick

Institut of Mouvement SciencesMarseille, France

Mr. Richardson

I utilize an example for better comprehension.

In gait analysis, because of different movement linear and angular velocity in different segment for optimum cut off frequency, according to the winter, (1) researchers should apply residual analysis for each of the markers attached to the body. The only problem of such process is time consuming, so there is another way for finding optimum cut off frequency is classifying the markers into segments which means, (2) utilizing one marker of each segment as reference for finding the optimum cut off frequency. For example in the plugin gait marker placement we have three markers on the ankle segment which we can use one of the markers as apply residual analysis for obtaining the appropriate cut off and apply to the other marker of the the ankle segment as well (not the other segments). (3) Another approach is utilizing one cut off frequency for all of the markers. In this process we should apply the residual analysis for at least two of markers, one with lesser velocity and the other with maximum velocity of the trial and then average them to find the most suitable cut off frequency.

I should mention that from approach 1 to approach 3, the diminution of accuracy and precision will be seen, and should analyzed, which the differences will not statistically significant. In the gait example winter find the optimum cut off as 6 Hz for all of the markers. So the quota of accuracy and precision you need is the key to find the way that detects the optimum cut off frequency.

Hope that it helps you.

Yours truly

Alighanbari

You can use vision.OpticalFlow object.

Hi Paul, 

Indeed, as Sina points out, a lot of measures are mentioned in my thesis. However, most of that is also published, and the best article to get you started would be this one :"Assessing the stability of human locomotion: a review of current measures.

Bruijn SM, Meijer OG, Beek PJ, van Dieën JH.

J R Soc Interface. 2013 Mar 20;10(83):20120999.". It's also online here on researchgate. In it, you will read that it's also possible to use short data series for these stability analysis. For this, I would highly recommend the work of Kim van Schooten. 

Best,

Sjoerd

Hi jacobus

sorry for the delay of my answer.

I never compare plugin gait and opensim model. I should do it.

You have to know that kinematic calculations of these models are different.

opensim considers a-priori definite joint axes and uses an inverse kinematic optimization (Lu, T., & O’connor, J. (1999). Bone position estimation from skin marker co-ordinates using global optimisation with joint constraints. Journal of Biomechanics, 32, 129–134.)

Plugin gait only considers ball joint at all articulation and extracts cardanic angles.

No model can be a golden approach !

regards

Fabien

You can try BioImageXD (http://www.bioimagexd.net/).

Hope it work...............

I think its possible.

First you should find and fix the "gaps" for the specific marker with a fourth order polynomium. This works well for gaps until 6 to 7 frames. If your frame rate is to high, you probably have more missing frames, but if the suggested path looks well, you may accept it.

You should try to set lower frame rates (100 Hz e.g.) and try this option.

Good luck with your work.

Dear David,

I'm not user of Simi Motion System, but I worked with other video analysis system. First you need to use minimum two video cameras if you want to create 3d area/angles and usually some 'calibration kit'.

I recommend KWON website ( http://www.kwon3d.com/ section THEORIES ) that contains a lot of the basic definition of the transition from 2d to 3d and transformation methods. Also you can find useful material from the ISB website

( http://isbweb.org/software/movanal.html ).

Best regards,

Greg (Grzegorz)