Implementation related to the paper "Analysis of deep complex-valued convolutional neural networks for MRI reconstruction and phase-focused applications" by Elizabeth K. Cole et. al; Toolbox for complex-valued convolution and activation functions using an unrolled architecture.
Thanks for your impressive work! Currently I also want to reproduce the complex unet in my dataset. But according to my reproduce results, I find the complex unet performance worse than real Unet. the best performance complex unet is same with input undersampled image. That's the problem. Would please share your reproduced complex Unet ? Thanks !
Is there a way to output the .pb file while training the model or if I use the freeze graph method what is the name of the input node and output node? May you please help me with this. I have tried to freeze the model but I do not know the name of the node I need to output which will be needed for inference.
I am trying to reproduce your results using the code and data provided.
The code runs without errors, however the output is NAN values. Loss function stays constant instead of decreasing.
Bart reconstruction seems to work fine.
Could it by a version issue? I am using Python 3.7, in combination with your requirements file. Hardware: nvidia rtx 3090, 24gb ram. Some packages were not listed in the requirements so I tried to find one that works. E.g. "gast" (version 2.2), skiimage. I have attached the requirements that I am using, maybe you can comment if that is correct.
Do you have any idea where to look for the error?
I would like to use the code later for preclinical MRI data (single coil). Could you comment on the following
Do I need to normalize the data or do any other preprocessing (right now it is complex k-space data)
Should I train each region and image type separately (e.g. head T1, abdomen T2) or can I train them together?
In case image size is different can I zerofill the images to have the same matrix size or does the original MR data need to be the same matrix size.