simtk.org

Hi everyone,

Is there a way to transform (translate and rotate) the local coordinate system for bones/segments? I want to change some muscle attachment sites, but the muscle data I currently have is defined in a different reference system than the model I am using. Specifically, I am interested in changing around the scapular muscle attachments in the DSEM (model), but the muscle data I have are defined relative to the AC joint as the origin of the scapular coordinate system, which is different than the DSEM (which has the origin of the scapular axes more to the back of the acromion). Similarly, the XYZ directions are differently defined in the model then the data I have. I tried changing parameters in the .osim code of the model, but all it would do is change the location and orientation of the entire child bone relative to the parent bone -- but I want to keep the location/orientation of the child bone the same, just shift where the axes within that child bone is located. I have access to SIMM and tried doing it in that software as well, but can't seem to find the function where this change can be made.

Any help would be appreciated.

Thanks!

Dan

Hi Dan,

If you want to change the position of the bone origin which you seem to want to do, from my experience to do this you would need to redefine the geometries used in that OpenSim model. The display geometries are defined in the local frame of the body. If you are using a scaled generic model, my initial thought is this solution is a lot of work.

Having looked at something similar I think it would be easier to go the other way. By this I mean, redefine the local position of the muscle attachments to be coincident with the body frames used in the OpenSim model you are using. Assuming along with the muscle attachments you have some idea of the geometry of the bones and some relative landmarks to create these coordinate frames.

Hope this helps,

Bryce

Hi Bryce,

Thank you so much for your reply!

I did realize shortly after posting, and your reply confirmed it exactly for me, that the coordinates in the bone geometry file are all defined within it's own local coordinate system -- I was thrown off because some of the bone files don't have an actual origin (0,0,0) line, which just meant that their LCS is off the bone. I've started to go ahead and use this to identify the landmarks on the the bone and change the origin location by translating all the nodes defining the bone geometry (this part is pretty straightforward). I agree with you though, this is a lot of work since the muscle locations, mass center, inertial centers, wrapping object locations, all need to be re-defined in the model as a result... but one of my goals was to match everything to standardized recommendations in the literature and keep it consistent. I appreciate you offering the alternative solution and will definitely keep it in mind if this doesn't end up working out.

Thank you!

Dan

mulladm wrote: ↑

Mon Apr 23, 2018 1:43 pm

I was thrown off because some of the bone files don't have an actual origin (0,0,0) line, which just meant that their LCS is off the bone.

Hi Dann
what do you mean by "bone file"? cause I'm searching for such a thing in OpenSim to find the relation between my FEM bone & OpenSim bone to reach the proper scale factors.
thanks

Hi Yunus,

I believe the bone file I was referring to was the bone geometry (.vtp) file, which come with many of the models available on OpenSim. Here's a link that has some details, specifically "Q. What are .vtp, .stl, .obj files? ":

https://simtk-confluence.stanford.edu/d ... +Questions

Important to realize that the bone geometry files are mostly for visualization purposes.

Hope that helps!

Dan