Scaled Model Orientation

Provide easy-to-use, extensible software for modeling, simulating, controlling, and analyzing the neuromusculoskeletal system.
User avatar
Ayman Habib
Posts: 2248
Joined: Fri Apr 01, 2005 12:24 pm

RE: Scaled Model Orientation

Post by Ayman Habib » Mon Jun 07, 2010 10:57 am

J.J.,

One underlying assumption is that the model captures all the degrees of freedom exercised during the activity you're simulating/tracking.
Adding the joint at the pelvis is a good start, but you may also want to look into the knee and ankle.

If scaling/marker-placement is bad then later steps will be even more problematic. so I'd suggest you work more on scaling/marker-placement before going to IK. You may also consider changing the weights of markers (in marker-placement) as well.

-Ayman

User avatar
J.J. Mordang
Posts: 26
Joined: Mon Mar 22, 2010 7:04 am

RE: Scaled Model Orientation

Post by J.J. Mordang » Wed Jun 09, 2010 3:42 am

Hi Ayman,

Thank you very much for your time and help!

I got some progress in the scaling but it's still not as it should be. The virtual markers are shifted to much for my taste even though I set them to a weight value of 1000. For instance I put 2 markers on the 5th metatarsal (one proximal one distal) and set a weight value of a 1000 because the skin is very thin there such that the experimental markers are well defined. But after scaling the virtual markers are shifted al the way to the proximal phalanx. Also markers surrounding the knee for defining knee angle are shifted such that the marker which were beneath the joint are shifted above the knee.

I really don't know what the problem is in my scale setup.


Kind regards,

Jan-Jurre

User avatar
Ayman Habib
Posts: 2248
Joined: Fri Apr 01, 2005 12:24 pm

RE: Scaled Model Orientation

Post by Ayman Habib » Wed Jun 09, 2010 10:50 am

J.J.,

Please submit a bug report and attach all the required files to reproduce the issue and I'll take a look into it. Again I suspect that the model may not have all the degrees of freedom necessary to track the markers but I'll take a look to try to pinpoint the problem.

Best,
-Ayman

User avatar
Kristy Godoy
Posts: 23
Joined: Thu Feb 05, 2015 12:53 pm

Re: Scaled Model Orientation

Post by Kristy Godoy » Tue Dec 04, 2018 9:33 am

Hello,
I am having trouble with the scaling using the model gait2392.
We have rotated our data using the transform tool in Preview Experimental Data.

We also had to delete some markers that we do not use during the capture, and created one for C7 (parent torso).
Now when I try to scale the model every segment below pelvis results rotated over it own axis. I think the problem is on the fact that the pelvis on the model is linked to ground, while my data shows the body of the person over the origin of the lab's coordinate system, so I think that foots should be linked to ground, Am I correct?
Can somebody give me a hint, please, of what is going on and what I should do to overcome this issue?
Here is an image of the result.

I have been searching on the forum for a post with this issue, the closest that I found was this post.
I appreciate any help!
Attachments
scaledH.png
scaledH.png (201.45 KiB) Viewed 1352 times

User avatar
jimmy d
Posts: 1375
Joined: Thu Oct 04, 2007 11:51 pm

Re: Scaled Model Orientation

Post by jimmy d » Tue Dec 04, 2018 9:49 am

Seems like the pelvis_range isn't large enough in your model, you should increase that. You should also double/triple check that both the experimental marker labels and the corresponding model marker names are correct.

User avatar
Kristy Godoy
Posts: 23
Joined: Thu Feb 05, 2015 12:53 pm

Re: Scaled Model Orientation

Post by Kristy Godoy » Tue Dec 04, 2018 11:00 am

Hi James, thank you so much for your reply.
I've checked the names and that isn't the problem, also because when I enter the scale tool, the tool recognizes all of them on the Static Pose Weights tab.
We did not change much of the osim file. Just markers through the GUI, if I have to increase the pelvis-range, I supposed I'll have to do that on the osim file, on pelvis_tilt, pelvis_list and pelvis_rotation, is that correct?
I did that, I modified to a high number (-180,180), the three of them, also to (-5,5), but nothing changed.

Is it possible that if we deleted some "important" markers from segments (foot, shank, thigh), the model somehow lost the reference to position the segments correctly?

User avatar
jimmy d
Posts: 1375
Joined: Thu Oct 04, 2007 11:51 pm

Re: Scaled Model Orientation

Post by jimmy d » Tue Dec 04, 2018 12:02 pm

Can you post a zip with the;
- Model
- Setup File
- Experimental static trial

User avatar
Kristy Godoy
Posts: 23
Joined: Thu Feb 05, 2015 12:53 pm

Re: Scaled Model Orientation

Post by Kristy Godoy » Tue Dec 04, 2018 2:23 pm

Here is the zip. I included the raw static trial.
Thank you so much.
Attachments
Gait2392_Modified.zip
(463.59 KiB) Downloaded 99 times

User avatar
jimmy d
Posts: 1375
Joined: Thu Oct 04, 2007 11:51 pm

Re: Scaled Model Orientation

Post by jimmy d » Tue Dec 04, 2018 2:50 pm

Your scaling is wrong. You don't scale your Pelvis and it is too large, so when the markerplacer runs IK on the model, you get an incorrect solution. Your solution would improve if you decreased the size (mainly height) of the pelvis.

Overall, your marker set is very sparse and you are going to have a difficult time coming up with good ways to scale the geometry and register the markers.

User avatar
Kristy Godoy
Posts: 23
Joined: Thu Feb 05, 2015 12:53 pm

Re: Scaled Model Orientation

Post by Kristy Godoy » Wed Dec 05, 2018 1:32 am

Hi James, I understand.
Thank you for taking the time to verify the data.
One of my colleagues tried using measurements he took during the trial and the scaling is improving.
Thank you so much for your help!

POST REPLY