search

Sara04 / BRATS

Brain Tumor Image Segmentation
21 stars 10 forks source link

How to run ur modules for the BRATS dataset #1

Open poornasandur opened 7 years ago

poornasandur commented 7 years ago

Hi Sara, I would like to know the directory structure for the BRATS dataset .. I mean how do you actually run it .Please suggest some pointers on this.

poornasandur commented 7 years ago

I meant how did you arrange the 4 modalities for all the patients in this folder which you have mentioned in the config2.json file

/home/sara/Science/databases/BRATS

Please suggest some pointers on this.

Sara04 commented 7 years ago

Hi Poornachandra,

Thank you for your email. I am glad if it is helpful. I didn't put entire project as I would like to clean it and organize it a bit, I will probably finish today or tomorrow and I will update the README file with the instructions on how to run it.

Best regards,

Sara

On Mon, Sep 11, 2017 at 7:04 PM, Poornachandra Sandur < notifications@github.com> wrote:

Hi Sara, I would like to know the directory structure for the BRATS dataset .. I mean how do you actually run it .Please suggest some pointers on this.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/Sara04/BRATS/issues/1, or mute the thread https://github.com/notifications/unsubscribe-auth/AFGOsfkj-p_GfGkTStoVZ-BYQZyeM3TEks5shWf9gaJpZM4PTekX .

poornasandur commented 7 years ago

Hi Sara, Actually I am glad to say that I am learning a lot from your code thanks a lot for making it online, I tried to run it for my BRATS dataset locally on my system, but failed to understand the exact directory structure from which the 4 MRI modalites of each patient are read.

Bonsen commented 6 years ago

@poornasandur @Sara04 It seems that the download of 17th data is closed now, could u share the url?

poornasandur commented 6 years ago

According to the license agreement of MICCAI BRATS 17, the data cannot be shared.

On Mon, Nov 20, 2017 at 12:19 PM, Zhang Boheng notifications@github.com wrote:

@poornasandur https://github.com/poornasandur @Sara04 https://github.com/sara04 It seems that the download of 17th data is closed now, could u share the url?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/Sara04/BRATS/issues/1#issuecomment-345606200, or mute the thread https://github.com/notifications/unsubscribe-auth/AXsXKD-TwQOcuCrRW0HKjN9yZ-TIjJgCks5s4SD0gaJpZM4PTekX .

-- warm regards, Poornachandra S +91-9901913049

Bonsen commented 6 years ago

@poornasandur ....OK thanks

TigabuYaya commented 5 years ago

Hi Sara, thank you so much for making your codes available freely. When I run the code, unfortunately, I came across with this error. so, please can you help me? Traceback (most recent call last): File "train.py", line 7, in from create_modules_objects import create_modules_objects_from_config File "/content/drive/My Drive/BRATSmaster/python/create_modules_objects.py", line 14, in import modules.segmentators File "/content/drive/My Drive/BRATSmaster/python/modules/segmentators/init.py", line 2, in from .cnn1 import CnnBRATS1 ImportError: No module named cnn1

Sara04 commented 5 years ago

Hi Tigabu,

Thank you for the information. Yes, it was an error. Actually, I didn't push CnnBRATS1 model, so I've removed its import from init file.

Best, Sara

TigabuYaya commented 5 years ago

Dear Sara, Thank you for your help. Can you please make these issues clear for me? As shown blow, form config2.json and create_modules_objets.py, shouldn't the corresponding names and mother_classes on the same row be with exactly the same name???

from config2.json from create_modules_objects.py

"name":"PreprocessorBRATSBasic" 'mother_class': PreprocessorBRATS,

"name":"PatchExtractorBRATSForCNN12", 'mother_class': PatchExtractorBRATS,

"name":"CnnBRATS2", 'mother_class': SegmentatorBRATS,

"name":"PostprocessorBRATSForCNN12", 'mother_class': PostprocessorBRATS

On Mon, Oct 29, 2018 at 10:42 PM Sara Sedlar notifications@github.com wrote:

Hi Tigabu,

Thank you for the information. Yes, it was an error. Actually, I didn't push CnnBRATS1 model, so I've removed its import from init file.

Best, Sara

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/Sara04/BRATS/issues/1#issuecomment-434178881, or mute the thread https://github.com/notifications/unsubscribe-auth/Ae-j-F74mP4OsUNnHqcb_m4fmINaT5vvks5up-a1gaJpZM4PTekX .

Sara04 commented 5 years ago

Hi Tigabu,

No, they should not. In the configuration file you should put corresponding child class names, as they contain definitions of the methods that are only declared in the mother class.

Best,

Sara

Mamoona81 commented 5 years ago

@poornasandur @Sara04 Hi ,I m new to medical images and neural network. Here my some Queries. 1.how to did you opened the mha file?is there any simple way to load it ? 2.Where to paste this command ''python train.py -config config2.json -o output-directory-path'' 3.How to run when I don't have GPU in my System. My configuration: Python 3.7 anaconda ,Pytorch

poornasandur commented 5 years ago

Hi, You can load it using ITK or you can convert them to .nii.gz files using C3D (convert 3d) tool from itk-snap website and after converting to Nifti format you can read them using Nibabel or SimpleITK. .

On Wed, Feb 13, 2019, 7:38 PM misbah54 notifications@github.com wrote:

@poornasandur https://github.com/poornasandur @Sara04 https://github.com/Sara04 Hi Here my some Quries. 1.how to did you opened the mha file?is there any simple way to load it ? 2.Where to paste this command python train.py -config config2.json -o output-directory-path

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/Sara04/BRATS/issues/1#issuecomment-463211769, or mute the thread https://github.com/notifications/unsubscribe-auth/AXsXKNpgEVSTPpNSfJYLxURGyqUPmkj7ks5vNBxJgaJpZM4PTekX .

Sara04 commented 5 years ago

Hi,

  1. I think you can also use python library medpy to load mha files.
  2. This command should be run from the directory where train.py script is.
  3. You can run it without GPU by using tensorflow CPU only version. You cannot use pytorch with this code, because the model is created and trained with tensorflow. In the readme file I provided python and all library versions which I used in the project.

Best,

Sara

TigabuYaya commented 4 years ago

Hey,

I just signed the petition "CALL FOR ACTION!!! I STAND WITH DR. TEDROS ADHANOM" and wanted to see if you could help by adding your name.

Our goal is to reach 25,000 signatures and we need more support. You can read more and sign the petition here:

http://chng.it/Ppzpc9Vv6L

Thanks! Tigabu