[vtkusers] voxel's volume

[Laurent PAUL]

> Laurent PAUL wrote:
>> Hi users,
>>
>> I'm wondering how voxels are mapped in my VolumeData.
>> I've got an ImageData coming from Dicom. Voxel size is 0.74*0.74*2.7, I
>> mean SizeX*SizeY*SizeZ(SizeZ=sliceThickness in dicom tag).
>> But acquisiton had been made with a spacing between slices=1.3
>>
>> So What is the voxel's volume?
>> Since I'm working in 3D, I'm wondering how voxels are mapped when
>> acquisition is overlapped.
>> I think I should consider the "spacing between slices" when it is
>> smaller
>> than
>> slice thickness. Am I right?
>
> Laurent,
>
> 	This is definitely a question for DICOM people not viz people. Please
> reread the PS3.3 of the DICOM spec, you'll see that Slice Thickness is
> only the thickness of tissues where the signal is acquired. What you are
> looking for, really is Spacing Between Slices.
>
> 	As a side note using this tag is generally extremely dangerous, as you
> generally don't know if slices are joints, overlapping or disjoints. For
> a better implementation look at ITK-GDCM where we take two slices, the
> Image Orientation (Patient) and compute the zspacing from those.
>
> HTH
> Mathieu
>


Hi Mathieu,

I agree with you. As I want to measure a volume counting voxels, it's a
very important question.
In the case where images are joints or overlapping, I can use spacing
between slices.
In the case where images are disjoint, measure using Spacing between
slices won't be the exact volume since we didn't acquired the whole body
area. It will just be an estimation.
I want to distinguish this two possibilities as it concerns medical
processing.

Thank you!

Laurent.

P.S.: In the case where Image Orientation is 1/0/0/0/1/0 (purely axial),
Spacing between slices is equal to the difference between two succesive
"Image Position Patient" values?
With an arbitrary orientation, I have to compute this difference with
Image Orientation representating cosines, right?