In general, for the processing, the size difference would not matter, as the processing is agnostic to the specific size (and there is no rescaling of the size of objects anywhere).
In your case you just need to make sure you create a reference image that accommodates all your input meshes (independent of their size). Thus, best choose your largest mesh and choose a fine enough sampling to create the reference image.
Martin
Hi Martin,
The ends of my specimens are broken, so we are now “cutting” the ends off them using new segmentations in slicer.
Would I still need to create a reference image from the largest specimen, or will my segmentations be okay just run directly in SPHARM-PDM?
Sorry this is a bit confusing, there is a lot here which is new to me!
Secondly for the spherical mapping, Brechenbuehler 1995 says that the first and last vertices which are mapped are the furthest apart in the z axis, so this will be the first and last points in the z stack of the segmentation? Does this mean that you can imagine the centre of sphere to be the centre of the object?
Thank you so much for all your help and patience
I really appreciate it
Fiona
If you start with surfaces, rather than segmentation images, you have to first convert the surfaces to binary volumes/images. And for that process you need a reference image. So, yes, you still need it.
Re spherical mapping: those first & last vertices are just used for the initial mapping (basically the south and north pole of the initial spherical parametrization). That parametrization is then updated/optimized to establish an equal-area-ratio mapping. The centre of the sphere (or spherical parametrization is not at the center of the object), but rather it’s a unit sphere (the center is at 0,0,0).
Martin
Hi Martin,
Thank you very much for the clarification. I am finally producing really nice models and I am very happy with these. Thanks again for all of your help (and sorry for the slow reply!)
Fiona