Hi,
I did the following:
bet2 A betted-A
flirt -ref MNI152_T1_2mm_brain -in betted_A -omat A2MNI.mat -out A2MNI.hdr
fnirt --in=A --aff=A2MNI.mat --cout=fnirted_A --config=T1_2_MNI152_2mm
invwarp -v --debug --warp=fnirted_A --ref=A --out=invwarped_A
flirt and fnirt was completed with the output files but
invwarp got stuck forever at Min & Max Intensities are: 0 and 0 as shown below:
Assuming warps was relative
invwarp:: Size = (250,250,256,3)
invwarp:: ROI Size = (250,250,256,3)
invwarp:: Dims = (1,1,1,1)
invwarp:: Minimum and maximum intensities are: 0 and 0
betted Image A was previously bias-corrected using N3 so later a
non-corrected betted-A image was used but still leads to the same invwarp
problem.
Image A is 1x1x1 mm voxel and I tried with (1) MNI152_T1_1mm_brain as
reference and (2) using MNI152_T1_2mm_brain but resampling Image A to 2x2x2
voxel, but still got the same invwarp problem.
I also tried with MNI152lin initially instead of MNI152 but still have the
same problem. Furthermore, I am not sure how to create the .cnf for
MNI152lin (can't find the brain_mask_dil for MNI152lin_T1_1mm in
fsl/data/standard dir).
A fellow colleague told me that bet2 (not bet) removes the header
information, which may lead to problem in doing flirt.
scanheader betted-A produces:
open_header.c: 232: cannot read 32 bit/pixel data from file 1000024_n3ed.hdr
The image header indicates an unsupported number of bits per pixel.
(AIR_INFO_BITS_ERROR)
She had suggested that I save under the image betted-A under Analyze as a
unsigned-8 bit image. I tried this but still got the same invwarp problem.
When I opened the unbetted-A, I discovered that it is a signed 16-bit image
so I saved betted-A under Analyze as a signed 16-bit image. However, I still
got the same invwarp problem.
Now, I am at my wits' end as to what I can further do to solve this problem.
I greatly appreciate any help. Thank you very much.
|