I am segmenting skulls from the NDMDID PM-CT dataset.
I want to measure endocranial volume, or some measure of cranial capacity/brain size. What is best or easiest way to do this from the skull models?
Here is an example image (segmented using Amira Avizo, not 3DSlicer):
Do I need to create an endocast or segment the brain to measure the volume?
EDIT: I used SwissSkullStripper on head/neck thin bone CT files to retrieve a volume measurement (see segmentation statistics), is this fairly reliable?
It is very easy to do endocranial space segmentation in Slicer. You can install the SlicerMorph extension and give SegmentEndoCranium module a try. It uses another extension called WrapSurfaceSolidify, which you can also use directly. That gives you more control over the segmentation:
It is possible that your scans are very high resolution. For endocranial space segmentation (particularly to get volume of the region), you don’t need that high-resolution.
It is possible that defaults are not good for your dataset. For example if your data is 0.1 mm thick 3.0mm smoothing would be 30x30x30 voxels which would indeed take long and will be unnecessary.
if SegmentEndoCranium didn’t work well, I suggest going straight to WrapSurfaceSolidify extension, which SegmentEndoCranium uses under the hood. It exposes more parameters
I mean if swill skull stripper is giving you what you want and it is faster, sure.
I do a lot of endocranial volume segmentation. The slicermorph extension is not working properly , as you experienced, but the WarpSurface Solidity is perfect, very effective and quite fast. Just use those settings:
Thank you.
I am now using the NMDID provided ‘TRUE_BRAIN’ CT scan for each decedent, which is actually working very well in SegmentEndoCranium (I can also increase the smoothing kernel size). From there, I am getting expected endocast volume measurements in segment statistics, so thank you!
Tom
