Mesh commands

AIMS library and commands usage

Moderators: denghien, riviere

Post Reply
kathleenACE
Posts: 9
Joined: Fri Apr 17, 2009 10:28 pm

Mesh commands

Post by kathleenACE »

Dear BrainVisa experts,

I have been making meshes from brain masks using the AimsMeshBrain command. I recently tried the AimsMeshWhite command and found that the results looked better. Can you tell me how these algorithms work and/or what is the difference between them? Is it appropriate to use AimsMeshWhite on a hemisphere mask? Is there a reason that AimsMeshBrain is used by the pipeline?

I also would like to get surface area, but as a continuous map across the brain. Perhaps as a texture? Is there a way to see the area of each of the triangles in the mesh?

Any information or documentation that you could point me towards would be appreciated as I hope to show that BrainVisa can operate on pediatric cases better than other software with this algorithm.

Thank you,
Kathleen
User avatar
Jean-Francois Mangin
Posts: 337
Joined: Mon Mar 01, 2004 10:24 am
Location: Neurospin, CEA, France
Contact:

Re: Mesh commands

Post by Jean-Francois Mangin »

AimsMeshWhite is dedicated to the grey/white interface in order to yield a spherical topology.
Using it for something else could lead to unexpected results.

AimsMeshBrain yields nice brain mesh only if coupled with VipOpenFold

Which age range are you dealing with?
kathleenACE
Posts: 9
Joined: Fri Apr 17, 2009 10:28 pm

Re: Mesh commands

Post by kathleenACE »

I prefer not to use AimsMeshBrain with VipOpenFold as it seems to miss the outer pial edge of the brain on my subjects.
The attached images are (1) the brain mask and sucal map; (2) the mesh from the pipeline overlaid on the mask, (3) the mesh that I made from AimsMeshWhite overlaid on the mask.
The mesh generated by AimsMeshBrain misses voxels, especially in the frontal lobe, which gives erroneous surface area and cortical thickness measures; whereas the mesh generated by AimsMeshWhite goes through the edges of the voxels, giving seemingly accurate surface area and cortical thickness measures. What do you think?
brain_mask.jpg
brain_mask.jpg (184.09 KiB) Viewed 21618 times
AimsMeshBrain+OpenFold.jpg
AimsMeshBrain+OpenFold.jpg (171.11 KiB) Viewed 21621 times
AimsMeshWhite.jpg
AimsMeshWhite.jpg (177.58 KiB) Viewed 21622 times
My subjects are ages 1-3 years. This subject is 2 years old.
Visually the AimsMeshWhite version seems better to me. Do you see any drawbacks to processing my data in this manner?

Thanks,
Kathleen
User avatar
Jean-Francois Mangin
Posts: 337
Joined: Mon Mar 01, 2004 10:24 am
Location: Neurospin, CEA, France
Contact:

Re: Mesh commands

Post by Jean-Francois Mangin »

Yes you are right, This stuff was only done to do nice pictures and there is no security
when the data are not perfect like with small kids. If you use AimsMeshSmooth maybe you can
get something nicer from the result of AimsMeshWhite.

Now I have a tick I use for problematic data that should replace the VipOpenFold stuff
because it is save/ I just remove the skeleton from the brain mask before meshing the brain.
This should be done after using the current pipeline, then you retrigger the mesh computation.

VipSingleThreshold -i Lskeleton_SUBJECT.nii -t 19 -c b -o Lskel.nii
VipSingleThreshold -i Rskeleton_SUBJECT.nii -t 19 -c b -o Rskel.nii
VipExtEdge -i Lskel.nii -m 3Da -o Lskele.nii
VipExtEdge -i Rskel.nii -m 3Da -o Rskele.nii
VipMerge -i Lskel.nii -o Lskel.nii -m Lskele.nii -v 0
VipMerge -i Rskel.nii -o Rskel.nii -m Rskele.nii -v 0
VipMerge -i voronoi_SUBJECT.nii -m Lskel.nii -v 0 -o voronoi_SUBJECT.nii
VipMerge -i voronoi_SUBJECT.nii -m Rskel.nii -v 0 -o voronoi_SUBJECT.nii
rm Lskel.*
rm Rskel.*
rm Lskele.*
rm Rskele.*
kathleenACE
Posts: 9
Joined: Fri Apr 17, 2009 10:28 pm

Re: Mesh commands

Post by kathleenACE »

good. I did subtract the skeleton out in the mesh posted below. sounds like you agree that this is an acceptable method for improving the mesh.
I think I will stick with the result of AimsMeshWhite without smoothing, as I don't want to lose any voxels for my cortical thickness and surface area measures.

thank you,
Kathleen
User avatar
Jean-Francois Mangin
Posts: 337
Joined: Mon Mar 01, 2004 10:24 am
Location: Neurospin, CEA, France
Contact:

Re: Mesh commands

Post by Jean-Francois Mangin »

To be honest, substracting the skeleton is OK as long as it is only for visualization purpose. I have no opinion about the consequence for thickness study, our mesh were not design for that... for instance, I never found the time to apply some of my topological tricks to guaranty sphericity of the pial mesh.
Betina Vase
Posts: 11
Joined: Mon Sep 24, 2012 1:09 pm

Re: Mesh commands

Post by Betina Vase »

Hi
Is it true that a mesh extracted from a sulci can not be used in a study of the thickness of the sulci?? I was hoping to do so :|

("I have no opinion about the consequence for thickness study, our mesh were not design for that..")

Thanks!

\Betina :)
User avatar
riviere
Site Admin
Posts: 1361
Joined: Tue Jan 06, 2004 12:21 pm
Location: CEA NeuroSpin, Saint Aubin, France
Contact:

Re: Mesh commands

Post by riviere »

Hi,
The meshes representing the sulci (the colored sulci elements) are in no way connected to any reality of thickness, because they are only meshes of the CSF/GM skeleton, whicj is always 1 voxel thick.
The GM/CSF and GM/WM interfaces meshes might be useable for thickness, but we have not (at least not yet) checked and validated their accuracy for such an application. It depends on the quality of the GM segmentation (which is getting far better in the latest versions of BrainVisa/Morphologist but it's always subject to caution), and on the accuracy of the mesh built from this segmentation.
Denis
Betina Vase
Posts: 11
Joined: Mon Sep 24, 2012 1:09 pm

Re: Mesh commands

Post by Betina Vase »

Hi Denis
Thank you for the quick reply!

I see you point. I thought that the GM/CSF interface was found along the wall in the entire sulci, so the space in between represent the CSF in the sulci? If the sulci is only one voxel thick is that then only one wall (left or right) or is it the medial or?

Best regards
Betina
User avatar
riviere
Site Admin
Posts: 1361
Joined: Tue Jan 06, 2004 12:21 pm
Location: CEA NeuroSpin, Saint Aubin, France
Contact:

Re: Mesh commands

Post by riviere »

Hi Betina,
It's the center of the fold.
Denis
Post Reply