Model Scaling Error

Provide easy-to-use, extensible software for modeling, simulating, controlling, and analyzing the neuromusculoskeletal system.
POST REPLY
User avatar
Zheqi Hu
Posts: 24
Joined: Tue Feb 16, 2016 5:39 pm

Model Scaling Error

Post by Zheqi Hu » Fri Apr 22, 2016 4:06 am

Dear OpenSim

Hello,
I'm trying to scale model with trc file from vicon nexus. When I preview the experiment data in OpenSim, everything goes well. But when I select the trc file in scaling control panel and click run, the software stops working and it needs to be closed.

I tried another trc file with less markers, the software could work well. So there might be something wrong in the trc file.

What should I do now to deal with this problem?

Looking forward your reply.

Regards.

Zheqi

User avatar
Aaron Godfrey
Posts: 55
Joined: Tue Feb 16, 2016 12:34 pm

Re: Model Scaling Error

Post by Aaron Godfrey » Mon Apr 25, 2016 6:15 am

Have you tried looking at the "weights" tab? I've had problems where having any of the coordinates enabled will crash the program. Or, if it's somehow allowing you to run while you have a marker enabled that doesn't exist in EITHER the trace or the current model, that would probably crash it, as well.

User avatar
Zheqi Hu
Posts: 24
Joined: Tue Feb 16, 2016 5:39 pm

Re: Model Scaling Error

Post by Zheqi Hu » Tue Apr 26, 2016 7:46 am

godfreap wrote:Have you tried looking at the "weights" tab? I've had problems where having any of the coordinates enabled will crash the program. Or, if it's somehow allowing you to run while you have a marker enabled that doesn't exist in EITHER the trace or the current model, that would probably crash it, as well.
Thank you. I've solved this problem, the coordination of one marker was wrong.

But do you know how to change the model height during scaling the model? The data preview went well, but the scaled model got twisted because of the height difference between the example model and the experiment model.

User avatar
Aaron Godfrey
Posts: 55
Joined: Tue Feb 16, 2016 12:34 pm

Re: Model Scaling Error

Post by Aaron Godfrey » Tue Apr 26, 2016 10:37 am

The change in height is depending upon the static model you use in the scaling tool (the primary function of the scaling tool is to change the model's height & weight). In light of the lack of a real tutorial on how to use the scaling tool, I wrote a post about this a couple of weeks ago to help someone else having trouble with it. Take a look at it here, and I'll continue after, assuming you checked it out and seen the screenshots:

https://simtk.org/forums/viewtopic.php? ... 560#p16560

Brief Recap: Basically, scaling is done by having two models - the original (current) and the static pose you are scaling up to - with identical markers. You make measurement sets using the markers, by pairing them off - for example, RASI + RKNL and LASI + LKNL make up the shank in my model. Once your measurement sets are built, you assign a set to as many bodies as you can. The distance between markers in each measurement set determines the scaling factor that you see in the 'scale factors' tab. If your static pose has a model that is taller than the current model, all (or most) scale factors should be > 1. If it's shorter, they'll be < 1.

That is where I'm guessing your problem is. If you do not assign enough measurement sets, then it probably won't be able to faithfully scale the entire model, but rather, only bits and pieces. For example, if in the example I give in the above-cited posted, I only scaled the pelvis and the torso, the arms and the legs/thighs/shanks would look goofy because they did not scale with the other two body parts. So my advice: double check that you have an adequate number of measurement sets (follow my example, those worked well for me), and then make sure they are assigned to the appropriate bodies. Then you should see scaling factors appear in the appropriate tab, and when you run the tool, the new model should be visibly taller or shorter than the previous.

Does that answer your question, or did I entirely miss the point?

User avatar
Zheqi Hu
Posts: 24
Joined: Tue Feb 16, 2016 5:39 pm

Re: Model Scaling Error

Post by Zheqi Hu » Wed Apr 27, 2016 5:30 am

godfreap wrote:The change in height is depending upon the static model you use in the scaling tool (the primary function of the scaling tool is to change the model's height & weight). In light of the lack of a real tutorial on how to use the scaling tool, I wrote a post about this a couple of weeks ago to help someone else having trouble with it. Take a look at it here, and I'll continue after, assuming you checked it out and seen the screenshots:

https://simtk.org/forums/viewtopic.php? ... 560#p16560

Brief Recap: Basically, scaling is done by having two models - the original (current) and the static pose you are scaling up to - with identical markers. You make measurement sets using the markers, by pairing them off - for example, RASI + RKNL and LASI + LKNL make up the shank in my model. Once your measurement sets are built, you assign a set to as many bodies as you can. The distance between markers in each measurement set determines the scaling factor that you see in the 'scale factors' tab. If your static pose has a model that is taller than the current model, all (or most) scale factors should be > 1. If it's shorter, they'll be < 1.

That is where I'm guessing your problem is. If you do not assign enough measurement sets, then it probably won't be able to faithfully scale the entire model, but rather, only bits and pieces. For example, if in the example I give in the above-cited posted, I only scaled the pelvis and the torso, the arms and the legs/thighs/shanks would look goofy because they did not scale with the other two body parts. So my advice: double check that you have an adequate number of measurement sets (follow my example, those worked well for me), and then make sure they are assigned to the appropriate bodies. Then you should see scaling factors appear in the appropriate tab, and when you run the tool, the new model should be visibly taller or shorter than the previous.

Does that answer your question, or did I entirely miss the point?
Yes, that's exactly what I need, and I have solve the problem now.

Thank you very much!

POST REPLY