open-discussion > general questions
Showing 1-2 of 2 posts
Jul 9, 2015 12:07 PM | Marco P
general questions
Dear Beatriz and Martin,
Thanks again for sharing SPHARM, It has been very useful for my researches. I still have a couple of question:
- I now how to show phi and tetha values on the SPHARM surface. Is it possible to do all the way round and show x,y,z coordinates on the sphere to see how a surface maps to the sphere?
- I'd like to calculate the vertex normal of some brain structures to perform some thickness analysis. Is this reasonable for you? If yes, did you release any tool in SPHARM to perform this analysis?
Thanks again for your collaboration,
Marco
Thanks again for sharing SPHARM, It has been very useful for my researches. I still have a couple of question:
- I now how to show phi and tetha values on the SPHARM surface. Is it possible to do all the way round and show x,y,z coordinates on the sphere to see how a surface maps to the sphere?
- I'd like to calculate the vertex normal of some brain structures to perform some thickness analysis. Is this reasonable for you? If yes, did you release any tool in SPHARM to perform this analysis?
Thanks again for your collaboration,
Marco
Jul 9, 2015 05:07 PM | Martin Styner
RE: general questions
Hi Marco
Re mapping on the sphere: The sphere is just a surface like any other surface, so you can attach point data to it via MeshMath. So you would need to get the x,y,z coordinates first out of the vtk surface files. This possible (but cumbersome) as the vtk forward currently saved by our tools is the basic ASCII vtk format. Unfortunately there is no tool, as we did not think about doing that. But if you open the vtk file into an editor you can see that there is a POINTS section and all the data in that section are the x y z coordinates. Now, you just need to rearrange those coordinates such that they can be loaded via MeshMath (look at the phi or theta files for a template of how to do that).
Re surface normals: Computing normals would be quite straightforward, but we don't actually have a tool to do so (it is a simple function call in the vtk library). Question though: Do you want to look at the thickness of a brain structure such as the hippocampus? Using surface normals would not be a good way to do this. We do offer the computation of a "medial" axis (enable the medialMesh option in the ShapeAnalysisModule or ParaToSPHARMMeshCLP). This will allow you to compute the distance to that medial axis, which can be interpreted as a type of thickness. Those measurements are in the file _medialMeshPartialRadius.txt as well as in _surfSPHARMMedialMesh.vtk. Currently this thickness is not yet available at the surface locations of the procalign Mesh which is something we need to do still.
Martin
Re mapping on the sphere: The sphere is just a surface like any other surface, so you can attach point data to it via MeshMath. So you would need to get the x,y,z coordinates first out of the vtk surface files. This possible (but cumbersome) as the vtk forward currently saved by our tools is the basic ASCII vtk format. Unfortunately there is no tool, as we did not think about doing that. But if you open the vtk file into an editor you can see that there is a POINTS section and all the data in that section are the x y z coordinates. Now, you just need to rearrange those coordinates such that they can be loaded via MeshMath (look at the phi or theta files for a template of how to do that).
Re surface normals: Computing normals would be quite straightforward, but we don't actually have a tool to do so (it is a simple function call in the vtk library). Question though: Do you want to look at the thickness of a brain structure such as the hippocampus? Using surface normals would not be a good way to do this. We do offer the computation of a "medial" axis (enable the medialMesh option in the ShapeAnalysisModule or ParaToSPHARMMeshCLP). This will allow you to compute the distance to that medial axis, which can be interpreted as a type of thickness. Those measurements are in the file _medialMeshPartialRadius.txt as well as in _surfSPHARMMedialMesh.vtk. Currently this thickness is not yet available at the surface locations of the procalign Mesh which is something we need to do still.
Martin