Bone embedded Joint Axes
Peter Boettcher
However in contrast to the instructions given in XROMM wiki I have
problems setting up the axes because I do not have normal anatomy.
Investigating CCL deficient stiles in dogs I might have significant
external or internal rotation between both bones which I would not
detect if I use the axes tool as suggested, because I do not know at
which time during my animation the ZERO rotation/translation occurs.
This is why I use the following work around, which I hope is a valid
way:
I use the jAx tool as suggested and calculate a set of transformations/
rotations. I break up the connections as suggested when reusing an
Anatomical Axis System. This way I get two Axes, one for the proximal
and one for the distal segment. Now I position the proximal Axes in
relation to the anatomical axis system of the proximal segment and the
distal Axes in relation to the anatomical axis system of the distal
segment. Finally I hit SetJointAxesSystem and gets 3 translations and
3 rotations, describing the 6 DOF movement of the distal segment in
relation to the proximal segment, based on two bone embedded axis
systems.
Does that sound right?
Regards, Peter.
Stephen Gatesy
I work on limbs too and need to set a reference pose for my bones in
order to set the 0 values for translation and rotation. Dave Baier
has worked quite a bit on doing this as objectively and repeatably as
possible. We can talk about that more later if you want. He might
also chime in if I've misunderstood your question.
Here's what we are currently doing. Rather than starting with models
in CT space in Maya and then animating each bone and then setting
anatomical axes for the JCS, we create a reference scene that has the
elements in their 0 translation and 0 rotation relationship. Because
hip, knee, and ankle angles are set to 0, bones interpenetrate in ways
never seen in life. We've been getting the bone long axis from the
model's least inertial axis and flexion/extension axes using a Matlab
script created by Trey Crisco's group in orthopedic engineering at
Brown to fit a cylinder to human femoral condyles. (Miranda,D.L.,et
al.,Automatic determination of anatomical coordinate systems for three-
dimensional bone models of the isolated human knee.Journal of
Biomechanics (2010), doi:10.1016/j.jbiomech.2010.01.036)
Appropriately placed and oriented giantLocators are parented to each
bone model so that they move with the animated element. Then, as you
say, the relationship between proximal and distal can be used to
calculate the 6 DOF joint kinematics.
Our setup is a bit more complicated because we're doing other things
as well, but the basic setup is like this:
pelvis model
pelvis ACS giantLocator (for world pelvic translations and pitch,
roll, yaw)
hipP ACS giantLocator (proximal hip axis located on sphere fit to
acetabulum)
femur model
hipD ACS giantLocator (distal hip axis located on sphere fit to
femoral head- aligned to hipP in reference pose)
kneeP ACS giantLocator (proximal knee......)
tibiotarsus model
kneeD ACS....
ankleP ACS....
tarsometarsus modle
ankleD ACS.....
Does this make sense? Each pair of locators starts superimposed in
the reference pose and then deviate as bones are animated with the
AbsTforms data.
It seems you have P and D axes parented to the bone models; do you
care about zeroing out each variable to start?
For each XROMM sequence we begin with a fresh reference pose scene.
We key this at frame 0 so that it is always there (animation typically
begins at frame 1).
Hope this helps.
Steve
On Aug 2, 1:35 pm, Peter Boettcher <boettc...@kleintierklinik.uni-
Boettcher, PD Dr. Peter
Thank you very much for the detailed explanations, but I am not sure if I got it right. Do you position the bones in such a way that both axes systems overlap each other perfectly? This would mean that the bones are not aligned anatomically at time point zero - right? Maybe you could show me a screen shot of your setup which illustrates that? How do one key the actual setup as you suggested in your mail?
What did you mean by "It do you care about zeroing out each variable to start?"
Thank you very much - Peter.
------------------------------------
Peter Böttcher, Dr med vet habil, DipECVS
European Veterinary Specialist in Surgery
Fachtierarzt für Kleintierchirurgie
Klinik für Kleintiere
Universität Leipzig
An den Tierkliniken 23
D-04103 Leipzig (Germany)
Tel: +49-341-9738700
Fax: +49-341-9738799
email: boet...@kleintierklinik.uni-leipzig.de
________________________________
Hi Peter,
tarsometarsus modle
ankleD ACS.....
Hope this helps.
Steve
--
You received this message because you are subscribed to the Google Groups "XROMM" group.
To post to this group, send email to xr...@googlegroups.com.
To unsubscribe from this group, send email to xromm+un...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/xromm?hl=en.
Gatesy, Stephen
When you bring the bone models into Maya they will be in CT space if you got the models from CT. Each bone MODEL will have 0's for xyz translation and xyz rotation, right?
We move and rotate the models into a pose (what I'm calling the "reference" pose) at which each JOINT will have 0's for xyz translation and xyz rotation.
For the guineafowl knee joint, for example, we want flexion/extension to go from 0 to 180° as the limb goes from fully flexed (0°) to fully stretched out (180°). This is measured about the blue axis at the knee, which points laterally so that we get positive Z rotation being extension (using the right hand rule). The long axes of the femur (red, X axis) and tibiotarsus (also red, X axis) are both in the horizontal plane (they aren't parallel because the knee's flexion/extension axis is offset from the femoral shaft) at 0°.
We compare the world-space locations and orientations of the proximal and distal ACS locators, which are children of their respective bone models, to get the JCS data. In the reference pose, all the joints are "zeroed" so that we can make sense of a 45° value for knee extension.
If you have methods of determining the ACS of each bone independently (we do not), then you should be able to skip this reference pose procedure. I assume that most knee researchers will understand your motion data if they understand your ACS's. In comparative zoological and paleontological research, even the concepts of ACS and JCS are relatively poorly known or understood (hope I'm being fair to my colleagues, but I didn't get the full power of this until a few years ago).
steve
Boettcher, PD Dr. Peter
Thank you so much for the two pictures - but I didn't get it yet. But your email already helped me a lot. This is what I did:
1. Import my two 3D-models (femur and tibia in CT-space).
2. Create a reference axes using the jAx-tool.
3. Align the axes with the tibia in terms of a bone embedded joint coordinate system
4. Then align the femur with the axes system.
5. Key the current scene at frame zero.
6. Animate the bones using the import tool.
5. Calculate the 6 DOF parameters with SetJointAxes in the jAx tool.
Does that make sense? Because I want to measure cranio-caudal motion of the femur (x-axis) along the tibial plateau I have to align the x-axis with the tibial plateau. In consequence flexion-extension angles are not what one might be used to as dogs usually have a 25 degree negative slope.
Do you think there is a way to become a user of the cylinder-condyle fitting tool you mentioned?
Many thanks - Peter.
________________________________
steve
To unsubscribe from this group, send email to xromm+un...@googlegroups.com <mailto:xromm%2Bunsu...@googlegroups.com> .
For more options, visit this group at http://groups.google.com/group/xromm?hl=en.
--
You received this message because you are subscribed to the Google Groups "XROMM" group.
To post to this group, send email to xr...@googlegroups.com.
To unsubscribe from this group, send email to xromm+un...@googlegroups.com <mailto:xromm%2Bunsu...@googlegroups.com> .