Page 1 of 1

Mesh alignment

Posted: Wed 31 Mar 2021, 18:41
by srupp
I am running openEMS mainly in the MRI domain and would like to know if there is a best practice to align meshes during post-processing. I am post-processing mainly via Matlab. I noticed that B1 and SAR dimensions vary relative to each other (on same run). Additionally, B1p dimensions will change, between between slight configuration changes. I tried to correct for this via small fiducial markers, but would prefer an easier way. I am using the virtual family models and would ideally even like to mask out certain brain regions and compare via script. I assume the virtual family can't be exported for legal reasons, but at least my grid references (fiducials) might be for both B1 and SAR?

Re: Mesh alignment

Posted: Wed 31 Mar 2021, 21:42
by thorsten

I'm not really sure what you mean or what exactly the problem is, but the meshing will always have influence on the results and accuracy.
Thus make sure to keep it constant during your investigations and refine it until the results converge. Otherwise I don't really know how to answer your questions.


Re: Mesh alignment

Posted: Thu 01 Apr 2021, 01:15
by srupp
I hope my description is more clear this time: I mildly modified your MRI Loop Coil template ( ... oop_Coil.m).
Do goal is to move the coil around a subject/phantom to simulate a multi-channel coil. Any time I am changing the location of the coil, the H-mesh dimensions change slightly though. I would expect that my mesh and airbox extends are large enough to accommodate moving the coil without change of H-mesh extends. I added 8 fiducial boxes, each in a corner of the sensor box to stabilize gridding between runs. But that doesn't resolve it and I would like to know how I can force a 2 mm grid over my core region.
Additionally, SAR varies in size by one voxel within the same run from the H-mesh. I am trying to understand how these can all be forced to be static.
B1p: [153×160×164 double]
B1m: [153×160×164 double]
SAR: [153×161×165 double]
Most preferably, I would like to get a 3D matrix in Matlab of each run, where each cell has a material ID so that i can align and stretch if necessary compare datasets between runs.
Thank you for your input!