[MINC-users] Strange problem with mritoself

[Thomas Mansi]

Dear all,

I'm using mni_autoreg0.98r and I'm having a very strange problem. Maybe
you could help me.

I want to coregister a T2 axial MRI to a T1 MRI (intrasubject), by doing
the following:

$ mritoself -lsq9 -mi -far -clobber \
-model icbm_avg_152_t1_tal_lin_symmetric \
-modeldir /usr/local/mni/data/models/ \
-mask target -target_talxfm xfms/patient_t1_to_tal.xfm \
proc/patient_t2ax_clamp.mnc.gz \
proc/patient_t1_clamp.mnc.gz \
xfms/patient_t2ax_to_t1.xfm

However, though this command works with our newly acquired (and
isotropic) MRI (from the Siemens machine), I get the following error on
some older anisotropic images:

[mritoself] [mansi at Toshiba:/cygdrive/e/Documents/NOEL/temp/FCD-mspec]
[2006-03-28 19:44:45] /usr/local/mni/bin/minctracc -clobber
/tmp/mritoself_3804/patient_t2ax_clamp_crop.mnc
/tmp/mritoself_3804/patient_t1_clamp_crop.mnc
/tmp/mritoself_3804/patient_t2ax_to_t1_tmp1.xfm -est_center
-est_translations -lsq9 -mi -groups 256 -threshold 100 0 -model_mask
/tmp/mritoself_3804/patient_t1_clamp_mask.mnc -step 7.3 7.3 7.3 -simplex 10
COG of v1: -0.076698 28.486593 23.346405
COG of v2: 0.285103 16.833281 29.833879
   [trans] = 0.361800 -11.653313 6.487474
Only center & trans required, now returning from init_transformation
slice lim 17 0
row lim   32 0
col lim   32 0
thresh =  100.00000    0.00000
No voxels were found in volume 1 with value above threshold (100.000000).
Error in /usr/local/mni/bin/minctracc in file init_lattice.c, line 545
Cannot calculate size of volume 1
.mritoself: crashed while running minctracc (exit status=65280)

I don't really understand what's going on here. The source MRI are first
nu-corrected using N3 (nu_correct patient_t2ax.mnc patient_t2ax_nuc.mnc)
and then clamped (grey level intensities clamped between 0 and 100). Of
course, the target T1 MRI has already been coregistered to Talairach
space. Minctracc seems to invert the min/max intensity values ...

The problem happens on my Cygwin system (laptop with windows XP) but
also on Linux and SGI machines. It seems to be related to the input
images, since the program works very well with recent volumes and some
oldest ones, but fails with others... Did I miss something? Must I
upgrade my mni_autoreg?

Many thanks in advance for your help!

Best regards,
Thomas Mansi


	

	
		
___________________________________________________________________________ 
Nouveau : téléphonez moins cher avec Yahoo! Messenger ! Découvez les tarifs exceptionnels pour appeler la France et l'international.
Téléchargez sur http://fr.messenger.yahoo.com