I am trying to acquire the coordinates of some bony landmarks of the femur and tibia. In the code they are defined in the global reference frame of the leg but I would like them in the ISB coordinate system. Luckily the rotation matrix for the ISB coordinate systems is given as well.
The problem arose when I used this rotation matrix to produce the ISB coordinates of the landmarks. The femoral epicondyles should lie on the plane spanned by the y- and z-axis of the ISB coordinate system. However this is not the case. Also the x- and z-axis seem to be interchanged.
This moved me to the definition of the rotation matrix, especially the RotMat function where three points are used as input. Via my supervisor I got an source where a similar function was used. I adapted it to the description given in the reference manual of AnyBody.
The result is a rotation matrix equal to that computed by AnyBody if it is first transposed and multiplied by minus one.
With my own definition of the RotMat function I managed to get coordinates that are more credible but still not entirely as they should be according to the ISB standard. For this I used the updated value of the medial femoral epicondyle instead of the standard which is not updated after running the initial conditions.
I included a PDF with my findings, this includes the formula I thought the RotMat function would be.
So to summarize;
- I think the wrong third point is used in the RotMat function to define the ISB coordinate system.
- I would like to know what this RotMat function looks like when using the three point coordinates.