Scaling Issues

[Kyler Howard]

I am working on a research project for my university over the summer where we are analyzing the spike motion of volleyball players using OpenSim. Unfortunately I have run into some issues when attempting the Inverse Kinematics myself, but I think watching the April 2017 webinar about scaling pointed me in the right direction. After watching it though, I did have some questions come up that were not addressed in the webinar.

For this project I am continuing research from a previous student and using their recorded data. I am using the MoBL_ARMS model, and unfortunately we only have one marker on the elbow, so I'm coming up with issues trying to find the X & Z scaling factors for the humerus, and having some issues with the thorax. I'm trying to add virtual markers into OpenSim and measure using those. With that, a few question have come up.

Firstly, how do I use virtual markers to make measurement sets if there isn't a corresponding marker in the experimental markers? Is that even possible?

With the measurement set's not working, I'm instead manually making scale factors. I can easily find the lengths in the model by making virtual markers, but without extra markers on the subject, I can't find the same lengths of the subject. I'm using Mokka to see the subject's experimental markers and measure between them. Mokka allows me to create new markers between two existing markers, and using that I am approximating the same virtual marker positions the best I can. Do you know a way to retroactively find those measurements, or if the original data didn't have them, are we just stuck? Do you also have any recommendations for different programs to measure these lengths?

Any recommendations for scaling, other models, inverse kinematics, or otherwise would be greatly appreciated.

Thank you in advance for your help,
Kyler Howard
George Fox University Student

[Michael Asmussen]

Hi Kyler,

I can offer some input on your questions.

Firstly, how do I use virtual markers to make measurement sets if there isn't a corresponding marker in the experimental markers? Is that even possible?

You can generate virtual markers in another program. In the past, I have used Matlab to generate the virtual markers from the experimental markers and then re-created a trc file with these new markers that can be used for scaling in OpenSim.

With the measurement set's not working, I'm instead manually making scale factors. I can easily find the lengths in the model by making virtual markers, but without extra markers on the subject, I can't find the same lengths of the subject. I'm using Mokka to see the subject's experimental markers and measure between them. Mokka allows me to create new markers between two existing markers, and using that I am approximating the same virtual marker positions the best I can. Do you know a way to retroactively find those measurements, or if the original data didn't have them, are we just stuck? Do you also have any recommendations for different programs to measure these lengths?

I don't know what exact measurements you are trying to determine with your method, but maybe you could rely on some regression models to scale these lengths based on other anthropometrics you have of the participants (e.g., height).

I hope this is helpful,

Mike

[jimmy d]

This video may also be useful;
https://www.youtube.com/watch?v=ZG7wzvQC6eU&t=108s