TVFromEigenSystem errors

[siddys]

Hi Gary,
I am trying to convert from FSL output through
TVFromEigenSystem, with the ultimate
objective being to generate the .mha file (pdrgb). However, i keep
getting these errors at when i
try to execute TVFromEigenSystem

terminate called after throwing an instance of 'std::out_of_range'
what(): basic_string::substr
Abort

I would like to add that i have been able to convert and get to the
mha file before using this very binary, checked it on ITK snap, and it
worked fine. I really do not know what is happenning now. Could there
be some environment variable that i forgot to set? DTITK_ROOT is
property exported. ldd shows no missing
libraries. I am running a bash script to convert a bunch of files, so
this is called from inside a script
which sets all the necessary variables and path.

best regards,
sid

[siddys]

Hi again,
Just another doubt.. I was checking the units of
diffusivity, and was wondering,
if at all i have to apply the scaling factor, at what stage of
processing should i multiply
the volumes? the DWIs-per direction images? should the b0 also be
corrected, if
this is the case?
thanks,
sid.

[Gary]

The error is likely due to library incompatibility. I have tested the
most recent binary release on the following dataset
http://www.seas.upenn.edu/~huiz/upload/dti_fsl.tar.gz . It works
fine. The command I used is: TVFromEigenSystem -basename dti -type
FSL . Try it yourself. If you see the same error, it must be a
library issue. Check if you have the 32-bit backward-compatible
libraries (glibc, libstdc++, ...) installed. If you only have the
current 32-bit libraries of these libraries installed, you are likely
going to have problems.

Gary

[Gary]

Adjusting diffusivity unit only affects registration. It shouldn't
matter here. Neither does it have anything to do with DWI (including
b0) images. You use your favorite tensor reconstruction program to get
the tensor volumes from DWI images. After that, you adjust the
diffusivity of the tensor volumes to be consistent with what is used
in DTI-TK.

Gary

[siddys]

Hi Gary,
Thanks for the reply. I tried with the data set that you
provided,
but now i get a different error:
reading the buffer ... ** ERROR (nifti_image_read): failed to find
header file for '_lambda1.nii.gz'

Since the files that were in the tarball were .img/hdr files, i
converted them to nii.gz, before i
ran TVFromEigenSystem. In any case (img or nii), the program fails
while loading lambda1 .

Also, can you please explain what the following means, quoted from
your previous mail

"After that, you adjust the
diffusivity of the tensor volumes to be consistent with what is used
in DTI-TK."

do i have to scale the L{1,2,3} V{1,2,3}, FA, MD etc??

thanks,
sid.

[Gary]

The good news is that the binary appears to work as expected. So the
program will work if you use it correctly. See my first reply and
make sure you have used exactly the same command I wrote. You can
read more about it in the documentation (http://groups.google.com/
group/dtitk/web/data-format)

As to the diffusivity adjustment, it means to adjust "diffusivity".
FA, V{1,2,3} are not diffusivity measures. FA is essential some ratio
of diffusivity measures and V{1,2,3} are just direction vectors. L
{1,2,3} are diffusivity measures. MD is a function of L{1,2,3}. So
in principle, you can multiply L{1,2,3} to adjust diffusivity unit.
However, it is easier to first convert to the NIfTI formatted tensor
format. Then use the command that I described in my documentation to
apply the adjustment.

Gary

[siddys]

Hi gary,
I had used lower case FSL. Now, it works for your data, but
i still get the same error for my data. same program... what else
can go wrong??

Thanks for the explanation about the diffusivity adjustment.

[Gary]

I don't know. Can you upload your data to some web server?

Gary