ginobilinie / medsynthesis Goto Github PK

I am very interested in your work, I now have some doubts：
(1) In your paper, the image size is 153X193X50, and The image size I used is 240X240X155, will the image size affect the result?
(2) In your paper, the MR size is 32X32X32, the CT size is 16X16X16, the image MR size obtained in the code is 64X64X64, the CT size is 48X48X48, but your note is THE MR size is 32X32X32, the CT size is 24X24X24, What dimension size should I take?

Hi all,

Please, I am struggling to understand how the latent space is used in your code.
I would like to know how MRI patches are related to the latent space to synthesize the pseudo-CT. If one can consider that the distribution used to initialize the filters can be considered as the latent space.

Thank you in advance for your understanding.

I‘m sorry to bother you. I want to reproduce your code.Howere,in the package I downloaded, it says that d_scale_model cannot be found. In this file: from d_scale_model import DScaleModel.Is there something wrong here？

I read the 3D code and didn't find a discriminator in it. So isn't the 3D model an adversarial network?

Hi，I have read your article, the method proposed is very good, but there are a lot of problems when running the code, I hope you can help me, I have got access to the ADNI database, but I don't know how to download the required data set, can you share the data set you use with me?

Hi, I just looked through your paper. I'm quite interested in the 3T to 7T part. It seems that the code published here is intended for MR-to-CT. Are you using the same code for 3T-to-7T training? Or do you have another version that you can share for the 3T-to-7T part for your paper?

we could not find the d_scale_model in fifth line in your d_model.py through your medsynthesis, could you give me some details? thanks!

Hello, first of all thank you for sharing. This is a great job. About GDL loss in 3dganversion/loss_functions/gdl3d_loss, If the shape obtained by the generator is (1,128,128,128,1), how can it become the gen_frames, gt_frames needed in the code. Besides, what is the value of alpha?
Best wishes！
Thank you！

Thank you very much for your excellent work. Now I plan to use MRI to transform CT. Do I need labeled MRI and CT data as input items to run the code? Looking forward to your reply.

I'm using your code for 3T to 7T conversion.
When i run the first program readMedImg4CaffeCropNie4SingleS.py to extract patches.
I'm getting this error
"IndexError: index 40000 is out of bounds for axis 0 with size 40000"?
Can you please help me in solving this.

HI,

I did not find CT images in ADNI dataset (DTI, MRI, PET, fMRI. but not CT), how can i get the aligned MRI and CT images.

In http://adni.loni.usc.edu/methods/documents/

"
1)The brain dataset was acquired from 16 subjects with both MRI and CT scans in the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database (see www.adniinfo.org for details). The MR images were acquired using a Siemens Triotim scanner, with the voxel size 1.2 × 1.2 ×1mm3 , TE 2.95 ms, TR 2300 ms, and flip angle 9◦. The CT images, with the voxel size 0.59 × 0.59 × 3 mm3 , were acquired on a Siemens Somatom scanner. A typical example of preprocessed CT and MR images is given in Fig. 1.

2. Our pelvic dataset consists of 22 subjects, each with MR and CT images. The spacings of CT and MR images are1.172×1.172×1mm3 and1×1×1mm3,respectively. In the training stage, we rigidly align the CT to MRI with FLIRT for each subject [49]. "

Thanks.

I 'm trying to generate CT from MRI with 3D vision, but the result doesn't look like CT. It's like a label map,no details.And the loss is already changing a little.Have you met this problem?

I could not find the usage of the Auto-Context Model in your code as descrided in your paper(Medical Image Synthesis with Context-Aware Generative Adversarial Networks)? Please give some details about the ACM in your code . thanks!

Thanks a lot for providing your code, I wonder which result is better? The 3D version or the 2D version?

I have downloaded the brainData in NIFTI format from IXI and put them in /raid/dongnie/brainData, but it reports an error:
FileNotFoundError: [WinError 3] The system could not find the specified path: '/raid/dongnie/brainData'
Do I need to process these data further?

System Information:

• Python Version: 2.7.12
• OS: Ubuntu 16.04 x64, GPU: Nvidia GeForce 940 MX
• TensorFlow Version: 1.4.0

Problem
Following is the traceback obtained on running python main.py

Traceback:

Traceback (most recent call last):
  File "main.py", line 51, in <module>
    tf.app.run()
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/platform/app.py", line 48, in run
    _sys.exit(main(_sys.argv[:1] + flags_passthrough))
  File "main.py", line 48, in main
    lr_step=FLAGS.lr_step,lam_lp=FLAGS.lam_lp, lam_gdl=FLAGS.lam_gdl, lam_adv=FLAGS.lam_adv, alpha=FLAGS.alpha)
  File "/home/aryan/Downloads/medSynthesis/g_model.py", line 58, in __init__
    self.build_model()
  File "/home/aryan/Downloads/medSynthesis/g_model.py", line 79, in build_model
    .minimize(self.d_loss, var_list=self.d_vars)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/training/optimizer.py", line 353, in minimize
    name=name)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/training/optimizer.py", line 474, in apply_gradients
    self._create_slots([_get_variable_for(v) for v in var_list])
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/training/adam.py", line 136, in _create_slots
    self._zeros_slot(v, "m", self._name)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/training/optimizer.py", line 796, in _zeros_slot
    named_slots[_var_key(var)] = slot_creator.create_zeros_slot(var, op_name)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/training/slot_creator.py", line 174, in create_zeros_slot
    colocate_with_primary=colocate_with_primary)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/training/slot_creator.py", line 148, in create_slot_with_initializer
    dtype)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/training/slot_creator.py", line 67, in _create_slot_var
    validate_shape=validate_shape)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/ops/variable_scope.py", line 1203, in get_variable
    constraint=constraint)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/ops/variable_scope.py", line 1092, in get_variable
    constraint=constraint)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/ops/variable_scope.py", line 425, in get_variable
    constraint=constraint)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/ops/variable_scope.py", line 394, in _true_getter
    use_resource=use_resource, constraint=constraint)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/ops/variable_scope.py", line 760, in _get_single_variable
    "reuse=tf.AUTO_REUSE in VarScope?" % name)
ValueError: Variable d_conv_dis_1_a/w/Adam/ does not exist, or was not created with tf.get_variable(). Did you mean to set reuse=tf.AUTO_REUSE in VarScope?

When I run the 3D FCN model, I get an error：
shape output G (?, 24, 24, 24, 1)
ValueError: Cannot feed value of shape (20, 1, 32, 32, 1, 5) for Tensor 'Placeholder:0', which has shape '(?, 32, 32, 32, 1)'

can anyone help?