Scale Problem

[Matthias Kunze]

Hi at all,

If I scale my model with an own markerset (and the other necessary files for scaling) I get some strange ankle angles after scaling the model (10° in both legs).
Can someone tell me how I could scale the model without this unrealistic change (because nobody start hopping with 10° dorsi flexed foot)?

Thanks for any help.

Regards,
Matthias

[Sam Hamner]

Hi Matthias,

I have seen this problem too, usually because it's difficult to place the virtual markers on the feet of the model in exactly the place as on your subject. To get around this, I usually take a photo of the sagittal view my subject during the static pose to estimate the ankle angle.

Then, when I use the Scale tool I set static pose weights for the coordinates ankle_angle_l and ankle_angle_r with a manual value. In a normal static standing pose, the ankle angle value is around 0 deg. I chose a low weight value, like 1 or 10, depending on how much error I suspect, so that it will still try to track the markers on the foot.

Here is a link to a screen shot from the scale tool showing where to set the Static Weight Pose for Coordinates in the Scale Tool: http://img847.imageshack.us/img847/5914 ... cposew.png

Best,
Sam

[Matthias Kunze]

Hi Sam,

thanks for your answer.
I think now it seems to be more realistic if I change the static weights of the coordinates (ankle_angle_l/r).
But I still have a question about the settings of the marker weights:

-I create measurement based scale factors (Screen shot: http://img853.imageshack.us/i/measurementset.png/)

- I set all my marker static weights to values to 1.0 (Screen shot: http://img29.imageshack.us/i/staticweightsmarkers.png/)
Is this the right way and is the scaled model now adjusted to my subjects anthropometric data?
Or should I set more important markers to a higher weight value (the lower the suspected error...the higher the value, or not?)
Could you provide some own example files where the model will be adjusted the a subject?
(I think in the OpenSim-example files doesn't base on experimental subject values)

Thanks very much for your help.

Matthias

[Sam Hamner]

Hi Matthias,

Below are my responses to your questions:

Q: "Is this the right way and is the scaled model now adjusted to my subjects anthropometric data?" A: Yes, this sounds like the correct procedure to me....

Q: "Or should I set more important markers to a higher weight value."
A: Ah, yes, you may want to increase the weights for markers that you can place with more confidence (e.g., I am more confident about markers placed on the knee epicondyles than a marker placed on the big toe of a subject wearing shoes). I also usually have higher weights for markers on the pelvis (i.e., ASIS and PSIS markers).

You can find all the settings files for my simulation in the zip file here: https://simtk.org/frs/download.php?file_id=2512

Look under the folder "scale" for a file called something like "subject02_setup_scale.xml" Here is a screen shot of those settings loaded into the GUI tool: http://img101.imageshack.us/img101/3452 ... enshot.png

Best,
Sam