batch conversion of mri diffusion images

[frastello]

Hello dears ,
I want to reproduce with command lines the behavior of the images recovery module of Brainvisa, for a Series05 (Exam***Series005***_raw.tar) tar file generated by the old MRI at Orsay.
I brought back directly from the archive system this tar file (I could also have downloaded it from the images recovery module), which contains a single PFILE***.7 file.
My aim is to convert this PFILE into a .ima file (in the Brainvisa recovery module, I should select the "automatic reconstruction" action to do that well).
I tried the PtKGhecSignaPFFileToGisConverter command (on a linux ubuntu 64bit station), but the .ima file generated is unreadable into Anatomist afterwards (moreover, this .ima file is at least 5 times bigger than the original PFILE, but this is probably normal).
Thanks a lot for your advice in advance!
Hace a nice day,
Fabien

[denghien]

Hi,

In fact the Ptk command lines are not manage in our package.
Howerver, it seems you use a right way to convert your data. How did you find this command ? Does someone advise you to use it or did you find it in a script ?
Have you tried with another Series ?

Isa

[cpoupon]

Dear Fabrice,

A PFile is a raw GE data. The command to be used to reconstruct such a data is PtkGehcEpiReconstructor.
Normally, you don't have to run it by yourself, as the data storage client has the feature to do the reconstruction and run the adequate script according to the data. If you look at the python code of the data storage client for GE systems, you can find all the details about the options to be used for the reconstruction.
For your information, access to PTK source codes is restricted and not open to everyone like other libraries developed by NeuroSpin developers, such as aims, vip, brainvisa, .....
Best

Cyril

[cpoupon]

Another point: as a PFile is a raw data, it is a complex data containing MRI k-space information, and cannot be converter simply as an image. As a consequence, you cannot effectively open it in anatomist, but the PtkGehcSignaPFile2GisConverter has two options '-m' and '-p' to extract the magnitude and phase image. But again, except if you are developing your own reconstruction system of MRI data, it is useless to look at such images....
Best,

Cyril

[cmeunier]

Dear everyone,

I have to work on these data, i.e diffusion Exam***Series005***_raw.tar generated by the old MRI at Orsay, which contains a single PFILE***.7 file. I try the command PtkGehcEpiReconstructor, the .ima was readable by Anatomist but the image seems to be "damaged" and the number of volume doesn't corresponds to the number of directions (141 instead of 41+1 b0). Moreover, there is a lot of noise and an extremely poor contrast. Of course, even with the correct bValueAndOrientation.txt it is impossible to analyse such images... If I use the option -saveComplex the .ima is not readable by Anatomist. I also need to use phase correction method refscan otherwise it makes a segmentation error.

Does the reconstruction command can explain this problem ? I try to use several options but I am not very familiar with this process, is there a specific way to use PtkGehcEpiReconstructor in my case ?

Thanks a lot for your advice in advance !

Cécile

[denghien]

Hi,

I suppose that you know Fabien. It seems that he has already seen this with Cyril Poupon who works on Ptk library.
You might talk about it with Fabien.
I am sorry I don't have any other information.

Isa