Pickle?

[elamarie16]

Dear Laura, I see you used pickle as your format for the input file. How did you convert your MRI images which would be in DICOM format into pickle? We use the NifTi format and have been struggling to find a method to pickle a NifTi file and any help would be greatly appreciated.
Best

[laurahanu]

Hi, I used Matlab to convert them from DICOM to 3D matrices, save them as .mat files and then convert them to Numpy in python. I believe there are some already made functions you could find online to do it automatically if you have Matlab, however, if you don't, it might be a bit trickier. Having had a quick look, there seem to be options in python as well, such as: https://github.com/danishm/mritopng You will probably have to compromise on image size however.. Hope this helps! Best, Laura

…

On Wed, Jun 27, 2018 at 2:04 PM elamarie16 ***@***.***> wrote: Dear Laura, I see you used pickle as your format for the input file. How did you convert your MRI images which would be in DICOM format into pickle? We use the NifTi format and have been struggling to find a method to pickle a NifTi file and any help would be greatly appreciated. Best — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#2>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AfKM0-LnX3uTz5u7DXVcPOa3Khs7aOz_ks5uA4LPgaJpZM4U5rNO> .

[simonepenna]

Dear @laurahanu, I see in your code that you load the data in this way:

# LOAD DATA
data=np.load(inPath, mmap_mode='r').astype(floatX)

So you have a single file and inside there is all the dataset? What type of file is it? I have 1000+ MRI Images in .nii format what do you suggest to achieve and reproduce your dataset?

One more thing: which version of theano and lasagne did you use? Because i have problem with them.
Best