There are no plans to have any direct support for gzipped NIFTI files in
SPM. The reason for this is that SPM has been written so that it can
randomly access particular voxels of the data. Reading such a voxel
would require the whole volume before the voxel of interest to be
unzipped. The overheads would be enormous.
Best regards,
-John
-----Original Message-----
From: SPM (Statistical Parametric Mapping) [mailto:[log in to unmask]]
On Behalf Of Bas Neggers
Sent: Wednesday, January 31, 2007 1:56 PM
To: [log in to unmask]
Subject: [SPM] nii, .gz and spm5 on linux?
Dear list,
I am currently setting up spm5 based fMRI postprocessing on our system,
and was wondering whether it would be easy to make spm5 read nii.gz
files instead of nii files. gzipping images would save us ~60% of space,
and on linux it should in theory be possible to pipe the images through
gzip before they enter preprocessing/model estimation (I don't care
about Windows). Slowing would not be a big issue, space is more of a
limitation.
Is there anyone that already has some experience with doing this? I
guess it would be easiest in the spm5 code on a central location (e.g.,
function) where images are always passed through when they are read, no
matter whether for preprocessing, stats or display purposes. For writing
nii I'd have the same question.
I already read the post below on spm5b that no .gz can be read/created
using spm5(b), but some time has passed since. And there is mention that
it should be possible from within spm5's @nifti toolbox (which I am not
very familiar with)...
http://www.jiscmail.ac.uk/cgi-bin/wa.exe?A2=ind05&L=SPM&P=R475147&I=-3&X
=77AC725B600F74C2F3&Y
Thanks for any pointers!
Cheers,
Bas
--------------------------------------------------
Dr. S.F.W. Neggers
Division of Brain Research
Rudolf Magnus Institute for Neuroscience
Utrecht University Medical Center
Visiting : Heidelberglaan 100, 3584 CX Utrecht
Room A.00.1.24
Mail : Huispost A.01.126, P.O. Box
3508 GA Utrecht, the Netherlands
Tel : +31 (0)30 2503386
Fax : +31 (0)30 2505443
E-mail : [log in to unmask]
Web : http://www.fmri.nl/people/bas.html
--------------------------------------------------
|