surface extraction of interhemispheric space

[MichelODU]

Dear BrainVISA developers,

I was wondering if there is a way to robustly extract the interhemispheric space in brain data with BrainVISA. I'm seeing encouraging results with the sulcal labeling, which I'm using for sulcal separable meshing, and in some areas, the two hemispheres appear fused together due to some areas in the scan where the tissue appears fused that way.

With a model-based approach such as you use for sulci, this make the model more complete, as well as improve my meshing .

From Denis, I received info on how to extract sulci individual and convert them to vtkPolyData files. I don't see hemispheres amongst these vtk files.

Best wishes,

Michel

[MichelODU]

Hi folks,

please hold off for now. Most of my tribulations are due to my own morphological processing, which can be improved to better delineate the hemispheres.

Michel

[MichelODU]

Hi again,

it looks like my need for a means to get the interhemispheric fissure is still an issue. Is there something in BrainVisa that can provide it?

Cheers,

Michel

[MichelODU]

Geez, I'm doing this whole thread all by myself, both questions and answers .

Looking at the pipeline, it seems that the voronoi computation produces the hemispheres, as well as subcortical volume comprising the cerebellum. This may be something that I exploit... e.g. if there is space between voronoi nbrhood 1 and 2, this is interhemispheric fissure, and be careful with any smoothing in that area for producing a mesh.

What is the logic of the voronoi based on? Is there a paper on this alone?

Michel

[riviere]

Hi Michel,
The voronoi hemispheres separation logic is basically to erode sufficiently the WM to break it into 3, 4 or 5 pieces, and then to use the broken pieces as seeds in a voronoi diagram growing into the brain mask to reconstruct the hemispheres and cerebellum. There are additional (optional) tricks to help cutting the corpus callosum which sometimes resists the erosion.
Denis

[MichelODU]

Hi Denis,

thanks for your kind reply, and sorry for my belated one.

I'm interested in finding more about those tricks...

Cheers,

Michel

[riviere]

Hi Michael,
Basically the tricks consist in digging holes in the corpus callosum (at fixed locations in Talairach space) so as to make it look like a dotted line which will break easily during erosion.
Denis