Re: Nifti conversion

[Jonathan Jackson-2]

On Mon, 18 Feb 2008 16:18:06 -0500, Rajagopalan, Venkateswaran
<[hidden email]> wrote:

>Dear All,
>
>It will be of great help to me if you could clarify the following;
>
>1) I am using Image J to convert my dicom images to NIFTI format. When i
load the nifti formatted images into FSL (to do image registration) these
images appear inverted, i also tried the rotate option before saving them
into NIFTI format but it doesn't work.
>
What happens if you save as Analyze 7.5 instead of NIFTI?  NIFTI images are
different to the older Analyze 7.5 in that they store information about the
image orientation. ImageJ does not take this into account when displaying
the image, but FSL does. Both programs should display the same though, if
you use Analyze format. There are some tools included with FSL that allow
viewing / editing of the orientational information, eg. 'fslhd' and
'fslorient'.  
 
>2) While doing T2 calculation using MRI T2 calculations option (using dicom
images) 16 bit i am not able to see different structures within the brain
clearly whereas if i convert my dicom images to 8 bit then i am able to see
all the structures within the brain in my t2 maps.
Try converting to 'float' before doing the calculation, and converting back
to 16 bit signed (I'm assuming your data are signed) afterwards with a
plugin found here: http://rsb.info.nih.gov/ij/plugins/converter.html.

regards,
Jon

[Daniel Glen]

There are several programs available for conversion of DICOM to NIFTI  
that might be useful for you.

http://lcni.uoregon.edu/~jolinda/MRIConvert/
ftp://surfer.nmr.mgh.harvard.edu/transfer/outgoing/flat/greve/
mri_convert
http://www.sph.sc.edu/comd/rorden/mricron/dcm2nii.html

There are also the to3d and Dimon programs from our group that can do  
this. You can use the nifti_tool program to view and change header  
parameters.

http://afni.nimh.nih.gov

FSLView may display images upside-down from what you might expect,  
but labels the directions correctly. FSLView seems to work best with  
"radiological" orientation (RAI = Right to left, Anterior to  
posterior, and Inferior to superior) for the storage order.

I would guess the issue regarding the 16-bits versus 8-bits is a  
display scaling one that you should be able to verify using one of  
the other tools to test this out.

Regards,
Daniel Glen


On Feb 25, 2008, at 9:04 AM, Jonathan Jackson wrote:

> On Mon, 18 Feb 2008 16:18:06 -0500, Rajagopalan, Venkateswaran
> <[hidden email]> wrote:
>
>> Dear All,
>>
>> It will be of great help to me if you could clarify the following;
>>
>> 1) I am using Image J to convert my dicom images to NIFTI format.  
>> When i
> load the nifti formatted images into FSL (to do image registration)  
> these
> images appear inverted, i also tried the rotate option before  
> saving them
> into NIFTI format but it doesn't work.
>>
> What happens if you save as Analyze 7.5 instead of NIFTI?  NIFTI  
> images are
> different to the older Analyze 7.5 in that they store information  
> about the
> image orientation. ImageJ does not take this into account when  
> displaying
> the image, but FSL does. Both programs should display the same  
> though, if
> you use Analyze format. There are some tools included with FSL that  
> allow
> viewing / editing of the orientational information, eg. 'fslhd' and
> 'fslorient'.
>
>> 2) While doing T2 calculation using MRI T2 calculations option  
>> (using dicom
> images) 16 bit i am not able to see different structures within the  
> brain
> clearly whereas if i convert my dicom images to 8 bit then i am  
> able to see
> all the structures within the brain in my t2 maps.
> Try converting to 'float' before doing the calculation, and  
> converting back
> to 16 bit signed (I'm assuming your data are signed) afterwards with a
> plugin found here: http://rsb.info.nih.gov/ij/plugins/converter.html.
>
> regards,
> Jon