Hi,
You need to make sure all the images are registered together and
resampled to
have the same voxel dimensions and matrix size. You can do this with
FLIRT.
Choose the image with the best resolution and contrast to be the fixed
reference in all of these registrations.
Once you have done this you can run fast in multi-channel mode (set the
-S option to 5 in your case) and then you should get your output
segmentations.
There isn't any way to add information from manually identified labels
(vois) to
the method, but why do you want to? Is it to better initialize the
segmentation?
All the best,
Mark
On 30 Apr 2009, at 12:40, Moran Artzi wrote:
> Hi,
> First-many thanks for your assistance.
> I want to combine outputs from different tissue segmentation
> We have the following modalities
> FA& ADC maps, T1W(pre+post Gd), T2W, FLAIR
> In addition we have voi of none tissue types (defined by expert) of -
> lesion, edema, GM, WM based on those tissues
> Can we use those voi (defined tissue types) for supervised tissue
> classification?
>
> Thanks
>
> On Thu, Apr 30, 2009 at 1:42 PM, Mark Jenkinson
> <[log in to unmask]> wrote:
> Hi,
>
> I'm not quite sure what you are asking here.
> Do you want to know how to run multi-channel fast (using different
> MR modalities)
> or are you asking how to combine outputs from different tissue
> segmentations?
> If you can be more precise about what you want then hopefully we can
> help.
>
> All the best,
> Mark
>
>
>
>
> On 30 Apr 2009, at 11:24, Moran Artzi wrote:
>
> Yes it dose help many thanks,
> However my main purpose is to use our training data set
> (use values from vois of tumor edema normal appearing WM and GM...)
> for several modalities and I didn’t understand how should I
> integrate our data set (vector of values for each vois) in the FAST
> tool.
> Can you pleas help me with this issue?
>
> Thanks a lot
> Moran
>
>
> On Thu, Apr 30, 2009 at 9:44 AM, Mark Jenkinson
> <[log in to unmask]> wrote:
> Hi,
>
> The pveseg file contains the best hard segmentation that follows
> from the partial volume segmentation results.
> That is, each voxel contains one number, representing the tissue
> type with the largest partial volume fraction. In general I would
> not recommend using this file for any computations, as using the
> individual partial volume segmentation results (e.g. *_pve_1.nii.gz)
> allows more accurate assessment of the segmentation and
> quantification of volumes. However, the pveseg is useful for a
> quick visual assessment of the segmentation.
>
> The mixeltype file represents the classification of each voxel's
> tissue mixture. That is, voxels containing only one tissue type
> have a different mixeltype from that containing mixtures of
> two tissues, which is different again from those containing mixtures
> of all three tissues.
>
> Hope this helps.
> All the best,
> Mark
>
>
>
>
> On 30 Apr 2009, at 06:55, Moran Artzi wrote:
>
> Hi,
> Sorry for the basic but I'm new user in the FAST Segmentation tool,
> and I did
> not fully understood the meaning of the output files:
> _pveseg.nii.gz and _mixeltype.nii.gz.
> Can you pleas help me with this issue?
> Thanks
> Moran
>
>
>
|