Effect of partial morphing

[OmarBahig (Omar)]

@pgalibarov
@Hamed-Shayestehpour

I have some technical questions regarding the effect of partial morphing at the knee joint (morphing distal femur and proximal tibia only) on the personalization of the MS model:

For Knee alignment:

1. Does performing partial morphing compromise knee alignment? in other words, if we have a subject with Varus alignment, does partial morphing at the region of interest (distal femur and proximal tibia) capture the alignment?

2. In case partial morphing does not capture the alignment, does full bone morphing do? and which is the best way/ method to capture it?

For ref frames for knee kinematics:
3. In the case of partial morphing at the region of the interest, it is clear that the talus will be excluded from the morphing process. What does this compromise in the personalization?

4. Since scans are partial taken at the knee joint, it is also clear that the femur head and malleoli are not captured. Thus, to redefine reference frames according to Grood & Suntay method, can I use the generic model (TLEM 2.0) femur-hip and shank-ankle centers**?** What are alternatives?

AnyRefNode JointCoordinateSystem = { //To track TF joint
AnyVec3 O = 0.5*(.FemurMedialEpicondyleNode.sRel_us + .FemurLateralEpicondyleNode.sRel_us);
AnyVec3 EpicondylarAxis = .FemurMedialEpicondyleNode.sRel_us - .FemurLateralEpicondyleNode.sRel_us;
AnyVec3 MechAxis = Main.Studies.HumanModel.BodyModel.Left.Leg.Seg.Thigh.HipJoint.sRel - O;
AnyVec3 MLAxis = cross(cross(MechAxis, EpicondylarAxis), MechAxis);
AnyVec3 APAxis = cross(MechAxis, MLAxis);
AnyVec3 sRel_us = O;
AnyMat33 ARel_us = {APAxis/vnorm(APAxis), MechAxis/vnorm(MechAxis), MLAxis/vnorm(MLAxis)}';
DEF_REFNODE_CUSTOM_SCALING_1arg(sRel_us)
DEF_REFNODE_CUSTOM_SCALING_ARel(ARel_us)
AnyDrawRefFrame drw = {RGB = {0,1,0};ScaleXYZ=0.09*{1,1,1};};
};

5. What would be the effect of considering the centers in question 4. (in case it is fine to use them)?

Thank you for your response in advance.

Best regards,
Omar

[dschander (Divyaksh S. Chander)]

Hi Omar,

Here are some answers:

1. Yes, the partial morphing should capture the Varus alignment. The knee joint axis is made by using control points projected on to the standard knee axis. The projected points are passed through the scaling function, so they should capture the Varus alignment. In case you want to adjust the Varus alignment further, there is also a variable called VarusValgusRotation in Thigh.KneeJoint. But normally you should be able to see it through partial morphing, especially since you will morph both shank and thigh. Using the VarusValgusRotation variable only adjusts on the thigh and is, let's say, a cheap and dirty solution.

2. Partial morphing will do it.

3. This is a generalized question that is hard to answer. It really depends on what is your motivation for this work and what are you trying to study. But, talking about knee joint, not having the talus should not generally affect it.

4. This is an interesting problem. Basically, the joint axis is defined by the projected control points on the tibia and femur. However, the neutral position of the knee joint is determined by the AnatomicalFrame. So, the knee neutral position is defined when the vertical axes of the Anatomical Frames on thigh and shank are aligned. In case of the thigh, the anatomical frame depends on the HipJoint center, and in case of the shank, it depends on the malleoli (You can see the code in Seg.any script file, it's similar to what you have done). With the partial scans, you will probably introduce some uncertainty about the neutral position if you use the standard hip and ankle joint centers. The scale hip joint and malleoli will not be representative of your subject. But, also I am not sure if there is a better alternative. Do you know the knee angle when the scan was taken? Then, you can estimate the offset and correct for it. Another way could be to define the knee joint based on the articulating surfaces, but I don't have a method or literature sources to suggest you for the knee joint (I haven't searched for it, but you are welcome to ).

5. I think I explained the effect in the previous point. If you use the default joint with scaling, you risk getting the neutral angle a little bit off.

Best regards,
Dave