I have surface .stls of the medial condyle of the knee that I would like to analyze with a ssm. However, I seem to have all bad particles at all angles up to approximately 75 degrees. Do bad particles mean that I should not export and analyze statistically, or is it more of a recommendation to attempt to improve the model? I have limited my particles down from default 128 to both 64, 32 and 16 which gives a better overall mean SSM, but still not good particles.
The good/bad particles may or may not be applicable to your data. How do the PCA modes look in Studio when animating the sliders? Are particles crossing over each other, or do they appear well behaved?
The figure below is as close as the particles get to crossing over. For reference the angle for good/bad particles below is set to 87 degrees. I would say overall they are pretty well behaved. We ran a Parallel analysis and got 6 significant PCA modes out of it (out of 20 some if I remember right?).
This is just kind of a 2D surface? The good/bad particles may not apply. It’s based on comparing the surface normals for each of the shapes at the correspondence points, but if you don’t have an inside/outside, they may be flipped for different shapes.
Essentially yes its like a 2D surface with some curvature (See side view below). We were thinking due to using surfaces that we should not need to worry about good/bad particles.
Yeah, I don’t think this metric will be meaningful as the surface normals for each of your inputs may be pointing the opposite way even for well corresponding positions.
Perfect! Thanks for your help.