These terms are not only for softwares. Some of them are software useful for videoanalysis (APAS, Simi Motion, Dartfish, Kinovea). Some of them are used almost only with association with measurement systems (Vicon, Qualysis). The choice of the software depends also on the mesurement system.

I have experiences with:

APAS - software for videoanalysis. It has several modules: capture, transform, filtering, display, ... We used it for movement analysis mainly in sport (external conditions)

Vicon - it is probably most used system for kinematic analysis. These system is based on ifrared cameras. The main software are Vicon Nexus (for measurement and data processing) and Vicon Polygon (data presentations: graphs, tables, etc.). These softwares are used almost only by Vicon systems users.

Thank you very much for your answers!

Dear Béres, I agree with Zdenek since you are presenting different softwares, hardware and integrated systems...

- Video based software (with or without markers, difficult to get real 3D motion analysis): Simi Motion, Dartfish, Kinovea and APAS (which for me is the best option if you are willing to get some accurate data)

- Infrared systems: Vicon, Qualysis, BTS... These group are "global" solutions. From motion capture to reporting of data including processing. They are VERY ACCURATE but expensive and are mostly used on clinical settings and for research.

- Data processing: VISUAL3D, MOKKA. They are softwares used for analysis of biomechanical data. For me, visual is the best option for research since you build your own models and computations but requires some biomechanical knowledge. They both are ONLY for processing, so you need extra hardware+software for capture and tracking

- Opti-track: Capture and tracking software (and usually hardware). You still need some software for biomechanical analysis (as visual or mokka) if you need biomechanical reports.

I have experience with APAS (for outdors), Vicon (excellent for clinical analysis), BTS (ultra acurate and easy to use), Visual 3D (the best processor but not user-friendly), Optitrack (good accuracy of capture with very good prices) and Simi Motion (It really depends on what you need.. can be good if you do not have any other option...)

Hope I could solve some of your doubts...

Dear Ursula,

Thank you very much for your answer. I asked this question because I’m about to study the practical possibilities of 3D motion analysis. I mean that I would like to use it real time or as quick as possible in training. For example I would like to define the most appropriate knee, trunk angle in set position for a sprinter in the given sport form, in given situation. I don’t want to guess, I don’t want to “feel” it. We use APAS. We can use it in a fairly “fast” way, but not real time. I don’t want to work with the selected system in laboratory circumstances, I don’t want to disturb the athlete, I just want to put the cams to the demanded position, shoot, short analysis then came out with the result to tell it to the competitor anytime, anywhere. It seems that there is no a correct solution.

Visual 3D has some real time features... But requires some automatic tracking tools... If you already perform tracking with APAS this could be a solution..

Have you thought on Kinect? It is less accurate (although it seems that version 2 really improved this aspect..) but provides you real time analysis..

Maybe this could also help you.. I just copied the post from another line of questions...

Juan L. Garrido-Castro · 6.75 · 5.26 · University of Cordoba (Spain)

At the University of Cordoba, we developed the UCOTrack Motion Analysis System (www.ucotrack.es). It allows to use video from any source (DV cameras, mobile phone, High Speed Cameras, ...). and produces 3D kinematic data. Although it will be commercially availble in the future, we accept beta testers. We used it in measuring spinal mobility in patients, gait analysis, etc. We also used in sport performance and horses. It is easy to use and can work in outdoors. But there is only a Microsoft Windows version :-(

Dear Ursula,

Thank you very much for your help. I registered but unfortunately I did not find the software to download and to try it out. As I red, this solution could be very useful. Do you know something about the download possibilities?

Sándor

Beres,

Your goal is definitely an ambitious one! Generally speaking there is a tradeoff between the ease of use for the system and the accuracy that it will give you. So, a system that is portable, simple, and fast will not necessarily give you highly accurate measurements. If the goal is to provide instant feedback and you are measuring in the sagittal plane only, I think you might be able to use something like an iPhone or iPad video capture app that allows for angular measurements to be taken. There are several that exist. Be warned, however, these measures will be subject to large errors if you are 'out of plane'.

Hope that helps.

Dear Sean,

Thank you for the answer. There is a saying: measure with micrometer, mark with chalk, cut with axe – or something like that. Yes I think on the one hand that the punctuality is very important, but not in every case. I think that there is not a key issue to know the knee angle with half degree accuracy on a block start training. However, is it the 3D Vicon or Qualysis system measured movement (with 3, or 4 infra high speed cameras) in laboratory circumstances so precise when the reflective bulb marker attached on the knee is 1 or 2cms (or more) far from the anatomically ideal center of the joint? I don’t think so. The question is - what is precise? What do we think about preciseness? I think on the other hand it must be a compromise “level” with the given task. In other words – why is it accepted to publish studies in high ip factor journals shooted and measured with high speed 3D top softwares and why not accepted with a more simple and practicable one. Is it necessarily significantly precise? What is the absolute preciseness? Is it reachable?

Hello Béres,

I think you received a comprehensive overview on the different systems and software already. I have used many of them myself under different circumstances. So which system you choose will depend on that as well.

Most (Qualisys, Vicon, MA, ...) high res, upper price range, multi-camera systems have added the option for active filtering to their systems which allows recording in almost any environment/light conditions (we did it in sunshine on snow, with Qualisys in that particular case).

Simi-motion, we used that in snow sports as well, requiring a very tedious almost manual tracking in that application (a few years ago) is good if you think of getting high quality synched video recordings (APAS will do the same but I have only very limited experience with that - many years ago). Simi also added a markerless tracking option recently which is very promising for fast feedback applications you talked of.

We work (play and validate) a lot with kinect at the moment which may prove very powerful - again for certain activities as it is currently very slow. Not sure if the just released kinect2 which is much better in resolution is faster. you may find this in announcements of it on the net.

The discussion went on to precision and what Ursula said is giving you a good clue on this issue. When applying marker-based systems you find literature, mainly on gait analysis, about reliability and comparisons between markersets etc. which gets you to validity. How much does a marker misplacement affect your results. These effects can be huge, especially looking at one of you examples, sprint start (tried that myself on an amputee sprinter some years ago using Vicon). Most marker-based systems will struggle: a) at extreme joint angles the skin movement artifacts will be larger than during upright running and b) if the athlete's body is curled together and low to the ground many markers will become invisible to the cameras. Sure you can play with camera positioning or markersets but that may move you away again from your quick feedback idea.

So this is not specific to the system mentioned and I suppose that kinect or markerless systems will experience similar problems in such body positions.

The newest version of C-motion allows you to apply inverse kinematics which may improve the joint movements in such situations but it may also mess it up. I have only begun playing with that (in much more straight movements than sprint start).

You also find some literature on using optimization to gain more consistent kinematics (vicon comes with optimization models, AnyBody uses this, it can be implemented in openSIM, ... and most likely many other software) but this will increase analysis time and would need to be validated before applying it specific tasks....

So unfortunately I cannot provide you with a specific answer or a selection guideline of what you should select. But I hope that this input will help in making your final choice.

Dear Uwe,

Thank you very much for your extremely useful writing. As I read your lines an idea came into my mind. What about to develop a 3D movement analyzer system. We don’t need systems necessarily using cameras. What if we give a stretch suit to the subject loomed with special wires in it. These would be something like transmitters. The transmitted signs are the 3D positions of them (something like “GPS, on surface”). So it would be a telemetric system with a receiver. We can get all the data as precisely as it can be and right at the given moment. I don’t know how to develop it, I don’t know much about the background of how to define telemetrically some object position in the space without visual contact, but as I suppose it is already a time-honoured method.

Here is another idea (more problematic…). The same base. Stretch suit with some special heated wires or dots filmed with thermo cameras. It doesn’t work in competition circumstances…

That’s all for sci-fi.

I am very curious about you mentioned - Simi and the added a markerless tracking option. How does it work (pixel color based)? What does it exactly mean markerless?

How do you mean inverse kinematics?

Optimization (AnyBody) equal smoothing data?

Thank you again!

@Béres Sándor

I dont know about C-motion mentioned by Uwe, but generally inverse kinematics means a reversed calculation of the postitions of the joints by knowing the position of a more distal element. In a sprint start this could be the foot. It's generally more difficult the more degrees of freedom there are in the joints between.

Easy answer about inverse kinematics, hope that's what you were asking for.

Hi again,

Philip is correct about the definition of inverse kinematics which requires that you have an underlying model which connects segments by defined joints (classical C-motion approach was to track all segments independently in 6 DoF). An option called inverse kinematics is now included in C-motion while it is based on a Global Optimization approach (Lu and O’Connor, 1999) such that you can also use markers on the segments within the kinematic chain.

Simi motion has included markerless tracking (http://www.simi.com/en/products/movement-analysis/simi-motion-2d3d.html) and somehow combines silhouette identification and pattern matching which may relate to some of the work I was involved in previously (http://citr.auckland.ac.nz/techreports/2005/CITR-TR-164.pdf).

Anybody includes a kinematic optimization procedure which is documented in Andersen MS, Damsgaard M, MacWilliams B, Rasmussen J (2010) A

computationally efficient optimisation-based method for parameter identification

of kinematically determinate and over-determinate biomechanical systems.

Computer Methods in Biomechanics and Biomedical Engineering 13: 171–183.

In regard to the fast feedback you aim at the inertial sensor approach may be a good option but comes with many practical details to be solved - as Sean said above: ambitious...

Not sure whether it's still relevant or not: I was reading a paper on whole body inverse kinematic analysis as attached- It presents a three-dimensional (3D) whole body multi-segment model for inverse dynamics analysis based on VICON motion capture data processing.

http://www.jbiomech.com/article/S0021-9290(08)00285-6/abstract

Thank you very much Pascal!

Dear Pascal, Where and how can I get this software? Is it for free or is for money? Thank you!