Hi,
adding the bvals to my previously working, vector-only gradient file resulted in an error in
reconstructing the tensor (see below for a fix).
Here's the command (automatically generated from GUI):
dti_recon "/data/dti/data.nii.gz" "/data/dti/trackvis" -gm "/tmp/dtk_tmp1/matrices/gradient.txt" 1 -b 1000 -b0 auto -iop 1 0 0 0 1 0 -p 3 -sn 1 -ot nii.gz
Here's the readout and error:
Using b values provided in gradient table
Did not find image info file '/data/dti/data.info'. Default info being used.
Segmentation fault (core dumped)
Here's the top of the 35 line gradient file:
0, 0, 0, 0.65
0, 0, 0, 0.65
0, 0, 0, 0.65
0, 0, 0, 0.65
0, 0, 0, 0.65
-1, 0, 0, 1000.65
-0.17, 0.99, 0, 1000.65
0.11, 0.66, 0.74, 1002.08
-0.9, -0.42, -0.11, 998.01
0.17, -0.6, 0.78, 998.01
0.81, -0.39, 0.43, 1004.65
Interestingly if I delete the '1' after the "…gradient.txt" the tensor computes (and of course the -b 1000 -b0 are not required)
Is this a bug? Obviously this work around only works for the command line. It fails in the GUI.
PS: Is there anyway to use dti_recon with a binary mask rather than using the b0 image ?
Thanks,
Neil
Login 
Register 
Search 
Forum
Print this Post 
