marker set list?

[Vincent Tchiong]

Hello everyone! Currently I am a grad student in bioengineering at Temple doing research for a master's project, and am using this model to do some mocap. I was wondering if there is a document that lists the markers used and what body parts they align with. I was unable to get the Vicon cameras in our university's facilities to track all of the markers on the spinal clusters, so I might have to edit the model by removing some markers to match the pilot data I collected. I have never worked with Opensim or Qualisys to this degree, so I would appreciate any advice I can get!

-Vincent Tchiong
Temple University

[Jacob J. Banks]

I don't believe the markers in the model are the current set we use. I've attached a marker file that you can replace that set with that will coincide with what we are currently doing.

I've also included a CCC model that is necessary to reduce the degrees of freedom of the spine from 51 to 6. Otherwise, it is pretty difficult to have track every vertebrae independently (see this paper for additional logic: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8358679/).

That said, you can adjust the markerset to what you think is best for capturing you movement. It is pretty easy/straightforward how to customize the markerset in OpenSim to match where you want markers.
https://simtk-confluence.stanford.edu:8 ... ker+Editor

Ok? Let me know if this doesn't make sense. You may want to work your way through some of the tutorials on the OpenSim site (if you haven't) to get more familiar with OpenSim before starting with our more complex model.
Jake Banks

[Vincent Tchiong]

Thank you for the assistance, I appreciate it! I am in the process of editing the marker set to better reflect the capabilities of the facility I am working with. I have worked my way through the first three basic tutorials and some of the other examples, but I do not believe I have ran into a CCC model before - how would I use it to reduce the degrees of freedom in the main model?

-Vincent Tchiong
Temple University

[Jacob J. Banks]

Open up the model I sent (in OpenSim). You'll see that 15 of the 17 IVJ are missing. In place there are CCC (Coordinate Coupling Constraints) that couple the other joints to the remaining joints (based on established associations). These are necessary when you have more DoFs than you have markers (well, 'necessary' is a bit strong, as you can get reasonable results without them). They are also common in leg models for synching the patella kinematics to knee flexion.

Careful if your doing SOpt... You'll need to have 2 models (ideally) though. As you can't use the CCC model in SOpt... or at least you can, but it is inaccurate (it throws off the muscle moment arms). In addition to the paper I referenced last time, see my dissertation for how I/we have implemented them (not same model, but we do this with this model too) and what moment arms look like with/without CCCs:
https://scholarworks.umass.edu/dissertations_2/2091/
Though the entire doc is a page turner! Look at the Chapter 4 methods (figure 4.2.3 on page 88 is a good) and Appendix C (especially Figure C.2. on page 174) to the jest of things.

To implement in a scaled model, open the .osim file sin a text editor (recommend Notepad++). Find the CCC section and replace one with the other (we do this programmatically via a Matlab to OpenSim API).

Probably confused you more than you wanted/were asking with just marker stuff! But you should be aware of this. Otherwise, sounds like your on the right track with markers.

[Vincent Tchiong]

I am definitely making progress, I've found the CCC section (the ConstraintSet part?) and copied it into the main model file in NotePad++. As per your dissertation, I see how it can make the results much more accurate. I ran into some difficulties when scaling the model, especially in the measurements section, many of the body parts and muscles look quite... distorted. The official documentation on this topic on OpenSim does not offer much insight into how to create your own measurement sets, rather than just telling the user to scale parts in 2D or 3D depending on the body part. I plan to add virtual markers to try and follow their advice further, and then redo the measurements. Are there any tips I should know from how you created your set for scaling?

Also, I do not have a force plate in the facility I am using at the moment.

[Jacob J. Banks]

Here is an example of a scaling approach we use. While it won't work for you (unless you have the exact same markerset), it will give some guidance for how to scale things in multiple dimensions. Unfortunately, this is often more art than science. When in doubt, just scale to height or a segment length.

We do use/define our own joint centers via virtual markers (particularly for the L5/S1 and hip joints) based off from some literature regressions. This is 100% necessary though, if you have a decent marker set (and especially with a repeated measures design) you should be fine.

I also included some actuator files. The 'feetonly' one will be needed if you do not have FPs. Otherwise the model won't have anything to fight gravity. You'll also want one the pelvis (the base segment) doing a similar thing if your participants in any position other than standing (e.g., sitting). It's also just good practice to have very small torque actuators at every joint to help balance kinetics (there is some posted OpenSim code, or maybe it comes with the download of OpenSim, to construct this file).

Otherwise, send a pic of your model and your attach your scale setup file and I can see if anything jumps out at me.

Let me know if you have any other questions.
Jake Banks

[Vincent Tchiong]

Thank you so much, the measurement set file has really helped with the logic of creating my own! I'm not sure how to use the virtual markers when scaling though, OpenSim will not allow me to run the scale tool if I don't have recorded data of the virtual markers - but I assume this can be circumvented by using only the experimental markers from my data collection process.

I have attached a picture of the model and the scale setup file I edited, there are some issues but it looks much better than before. I do not have wrist markers or additional head markers, which I will need to add for the next iteration of this experiment.

[Jacob J. Banks]

Getting there!
Try locking all the spine and rib joints. That should get rid of the funky stuff (then unlock them after you scale). Otherwise, you also need to make sure your scaling every vertebrae the same way.
If your missing arm markers, just scale manually (easiest with a ratio of subj height / base model height).
Jake Banks

[Vincent Tchiong]

I ended up changing some of the markers as well as re-recording the data for the rest of the experiment, and eventually it worked out! I performed inverse kinematics analysis, as I could not get access to the facilities with force plates in time to properly work with inverse dynamics analysis. Wanted to say thank you for all the help you provided on this forum, and that I will be presenting my master's project over Zoom tomorrow. If you have the time to attend (which would be amazing) it will be at 1:00 - 1:30 PM, and I can email you the link.

-Thanks again,
Vincent Tchiong

[Jacob J. Banks]

Good deal. Congrats and good luck!
I might have time to attend... if that is ok with the school etc. Probably better to email the link though, so it isn't on a forum:
jbanks3@bidmc.harvard.edu
If I can't make it, good luck again!
Jake Banks