Hi,
You could do this, but the lesion voxels might mess up the mixture
modelling so it's probably best to delete the voxels before running
FAST.
Why don't you try it both ways and see which works better?
Steve.
On 17 Feb 2008, at 06:37, Ruchika Wadhwa wrote:
> Yeah, I thought of passing that image into FAST. However, I am using
> ANALYZE
> to manually trace the lesions on a T2 image that is registered on a T1
> image. So the image that I have is a transformed image on which
> lesions can
> be easily identified.
> I am not sure if this technique would work, so instead of passing the
> transformed image into fast, I was thinking of creating a binary
> map of
> that image with lesions having a value of zero and multiplying that
> with the
> *_seg.hdr image that fast generates and then segmenting that into
> gray,
> white and csf. However, I am not sure, first if this would be a valid
> approach and secondly, since fast generates the *_seg.hdr image and
> the
> partial volume estimates together, how could I modify it (fast) in a
> way
> that it would let me use a modified version of the *_seg.hdr image?
>
> Hope this makes sense. Thanks!
> -Ruchika.
>
> On Feb 16, 2008, at 11:57 PM, Steve Smith wrote:
>
> Hi - sure, you can do that - you can just use fslmaths to zero those
> voxels
> before passing the image into FAST.
> Cheers.
>
>
> On 16 Feb 2008, at 22:15, Ruchika Wadhwa wrote:
>
> Hi,
>
> I was wondering if I have a binary mask for each subject with the
> lesions
> manually traced, could I then use it as a weighted image (in FAST)
> in order
> to mask out the areas that are classified as lesions and then
> estimate gray
> and white matter volumes?
>
> Thanks,
> -Ruchika.
>
---------------------------------------------------------------------------
Stephen M. Smith, Professor of Biomedical Engineering
Associate Director, Oxford University FMRIB Centre
FMRIB, JR Hospital, Headington, Oxford OX3 9DU, UK
+44 (0) 1865 222726 (fax 222717)
[log in to unmask] http://www.fmrib.ox.ac.uk/~steve
---------------------------------------------------------------------------
|