Hi Falk,
can you give us a big picture of what you are trying to accomplish?
Bruce On Wed, 10 Nov 2010, "Falk Lüsebrink" wrote:
Hi Bruce,
I'm so sorry, I should have double-checked. You are completely right; mri_normalize does work perfectly fine. The conformation to 1mm isotropic is done in the last step of mri_nu_correct.mni.
In the last step of mri_nu_correct.mni the --conform flag is used, which shouldn't be used:
mri_convert ./tmp.mri_nu_correct.mni.32343/nu1.mnc nu.mgz --like orig.mgz --conform mri_convert ./tmp.mri_nu_correct.mni.32343/nu1.mnc nu.mgz --like orig.mgz --conform $Id: mri_convert.c,v 1.166.2.2 2010/08/10 19:11:50 greve Exp $ reading from ./tmp.mri_nu_correct.mni.32343/nu1.mnc... TR=0.00, TE=0.00, TI=0.00, flip angle=0.00 i_ras = (-1, 0, 2.12874e-08) j_ras = (3.72529e-08, 0, -1) k_ras = (0, 1, 0) Original Data has (0.7, 0.7, 0.7) mm size and (321, 321, 321) voxels. Data is conformed to 1 mm size and 256 voxels for all directions INFO: transform src into the like-volume: orig.mgz changing data type from float to uchar (noscale = 0)... MRIchangeType: Building histogram Reslicing using trilinear interpolation writing to nu.mgz...
Wed Nov 10 10:47:32 CET 2010 mri_nu_correct.mni done
Also attached is the recon-all.log and the output of mri_info of orig.mgz and nu.mgz.
Thanks again, Falk Lüsebrink
-------- Original-Nachricht --------
Datum: Tue, 9 Nov 2010 21:06:49 -0500 (EST) Von: Bruce Fischl fischl@nmr.mgh.harvard.edu An: "Falk Lüsebrink" falk.luese@gmx.net CC: freesurfer@nmr.mgh.harvard.edu Betreff: Re: [Freesurfer] -noconform flag of mri_normalize doesn't work
Hi Falk,
I just tried it and it seemed to work for me:
mri_info Gaab20070225a/mri/rawavg.mgz Volume information for Gaab20070225a/mri/rawavg.mgz type: MGH dimensions: 256 x 256 x 128 voxel sizes: 1.0000, 1.0000, 1.3300 type: SHORT (4) fov: 256.000 dof: 0 xstart: -128.0, xend: 128.0 ystart: -128.0, yend: 128.0 zstart: -85.1, zend: 85.1 TR: 2000.00 msec, TE: 3.44 msec, TI: 900.00 msec, flip angle: 9.00 degrees nframes: 1 PhEncDir: ROW ras xform present xform info: x_r = 0.5924, y_r = 0.0086, z_r = -0.8056, c_r = -9.9942 : x_a = -0.7780, y_a = -0.2537, z_a = -0.5747, c_a = 13.5548 : x_s = 0.2093, y_s = -0.9672, z_s = 0.1436, c_s = -22.2562
mri_normalize -noconform rawavg.mgz test.mgz not interpolating and embedding volume to be 256^3... reading from rawavg.mgz... normalizing image... MRInormInit: Talairach origin at (142, 110, 71) wsize 10, windows 14 above, 9 below max gradient 1.000 . . . mri_info test.mgz Volume information for test.mgz type: MGH dimensions: 256 x 256 x 128 voxel sizes: 1.0000, 1.0000, 1.3300 type: SHORT (4) fov: 256.000 dof: 0 xstart: -128.0, xend: 128.0 ystart: -128.0, yend: 128.0 zstart: -85.1, zend: 85.1 TR: 2000.00 msec, TE: 3.44 msec, TI: 900.00 msec, flip angle: 9.00 degrees nframes: 1 PhEncDir: UNKNOWN ras xform present xform info: x_r = 0.5924, y_r = 0.0086, z_r = -0.8056, c_r = -9.9942 : x_a = -0.7780, y_a = -0.2537, z_a = -0.5747, c_a = 13.5548 : x_s = 0.2093, y_s = -0.9672, z_s = 0.1436, c_s = -22.2562
On Tue, 9 Nov 2010, Falk Lüsebrink wrote:
Dear FreeSurfers,
The -noconform flag of mri_normalize doesn't seem to work properly in
the centos4_x86_64_stable-pub-v5.0.0 build. I have been trying to process a volume with a native isotropic resolution of 0.7mm using recon-ll including the -cm flag to conform the input to minimum. But during the normalization of mri_normalize the volume is conformed to an isotropic resolution of 1mm.
The flag is correctly used during mri_normalize, but it doesn't seem to
do anything. The recon-all log says:
"not interpolating and embedding volume to be 256^3..."
But the output volume T1.mgz is conformed to 256^3 and therefore to an
isotropic resolution of 1mm. This also happens if mri_normalize is used by hand using the -noconform flag.
A fix is highly appreciated.
Thanks in advance, Falk Lüsebrink