If you have a contrast enhanced image, you can use the level-set segmentation
to segment the brain vasculature. I'm not sure if there is anything different
about Gadolinium T1 MRI than the images that our lab uses, but you should be
able to segment them using one of the initialization types within the
"vmtklevelsetsegmentation" script.
To figure out what threshold values to choose, use the "i" button when the
image set has been displayed. Take note of the intensity values within the
vessel and in the background. Your lower limit should be just slightly below
the intensity values within the vessel itself. Since the MRI images don't
display bone, your upper threshold value will likely be set to "n" for no upper
limit. To segment the vessels, you'll likely want to take advantage of the
colliding fronts initialization type. Once a threshold value is chosen, you'll
be asked to place seed points. This is kind of a trial and error process. Your
seed point location will depend on the amount of noise in your image set and
gradient differences between the background and the enhanced vessels. You may
have to choose seed points closer to one another to capture vessels in areas
with low contrast between background and vessels.
If you just want a rough segmentation of the vessels, you can use the
isosurface function to generate a level set segmentation of the entire image
set with a given low end threshold value. Once you run the generated image file
through the vmtkmarchingcubes script to generate a surface file, you can apply
the vmtksurfaceconnectivity to eliminate small bits of the surface that aren't
continuous with the main vessel structure.
Here's some lines of script that you could use to generate your vessels:
vmtkimagereader -f dicom -d PathToDicomImageDirectory --pipe
vmtklevelsetsegmentation -ofile Outputpath/Outputfilename_LS.vti --pipe
vmtkmarchingcubes -ofile Outputpath/Outputfilename_LS_MC.vtp
--vmtksurfaceconnectivity -ofile Outputpath/Outputfilename_LS_MC_SC.vtp
Hope this helps,
Ben Dickerhoff
________________________________________
From: Tiziano D'Albis [tiziano.dal...@univ-rennes1.fr]
Sent: Friday, June 01, 2012 3:36 AM
To: vmtk-users@lists.sourceforge.net
Subject: [vmtk-users] segmentation of brain vessels from Gadolinium T1 MRI
Hi,
I am wondering if there is any chance to use VMTK to segment brain
vessels from Gadolinium T1 MRI.
I gave it a quick try using the level-set segmentation, but I had an
hard time finding seed points and/or thresholds.
Just want to know if you think this would be feasable.
Thanks
Tiziano D'Albis
------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
vmtk-users mailing list
vmtk-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/vmtk-users
------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
vmtk-users mailing list
vmtk-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/vmtk-users
