Hi Steve,
Thank you very much for your answers! It is a lots help.
I used avwstats to add up the lesion volume, like
%avwstats s8_brain_pve_0 -R -V -M
0.000000 0.999992 156251 411989.941406 0.370034
%avwstats s7_brain_pve_0 -R -V -M
0.000000 0.999997 151836 400348.828125 0.414307
It looks like that it gives a volume include all intensity from 0 to 0.999.
Can I give a lowthreshold and upthreshold to limit the intensity (like
>0.5), though I don't know if it is necessery.
I also used the medx to caculate volume of lesion image (*_brain_pve_0). It
is about
7.2576e06 mm*mm*mm for both time point
without giving threshold limit.
If I gave it a up and low threshold(1 0.5), then I get
151308mm*mm*mm and 134997 mm*mm*mm
for each of time point.
My purpose is to compare the accuracy between MEDx and FSL for our study.
I couldn't understand why they are so different between two method.
Am I missing somewhere?
Best,
Qingyan
>From: Stephen Smith <[log in to unmask]>
>Reply-To: FSL - FMRIB's Software Library <[log in to unmask]>
>To: [log in to unmask]
>Subject: Re: [FSL] Estimate Lesion volume
>Date: Thu, 16 Aug 2001 11:34:38 +0100
>
>Hi.
>
> > So I can't use sienax to caculate the brain volume.
>
>sure - you can, but by default, probably lesions will be included in this
>volume (which in fact most clinicians seem to think is the most sensible
>choice if it has to be one way or another).
>
> > Actually, I have gotten
> > 4 classes of segmentation brain image by FAST. But I don't know what I
>got
> > to do next.
>
>so - assuming that the <input>_seg.hdr image looks like you are getting
>correct segmentation, all you need to do now is estimate the volume of the
>approriate class. The most accurate way of doing this is to carry out
>partial volume modelling, for example:
>
>fast -c 4 -e -ov input_brain
>
>you can then add up the partial volume voxels by (for example for class 0):
>
>avwstats structural_brain_pve_0 -V -M
>353422 706844.000000 0.389197
>
>353422 gives the number of partial volume voxels
>0.389197 gives the mean value for these voxels
>
>so multiplying these two numbers together gives the total volume for this
>class.
>
>Thanks, Steve.
QingYan [log in to unmask]
(425)391-8251QingYan Guan
[log in to unmask]
(425)391-8251
_________________________________________________________________
Get your FREE download of MSN Explorer at http://explorer.msn.com/intl.asp
|