About code

[zwq1111]

Thanks for the code you provided. But I meet some problems when I run the code, I want to know whether the code is entire or not. If it is entire, I am so sorry to bother you. If it is not, I would like you to post the complete code. Thank you very much again!

[zhiyuns]

We are sorry for the potential problems with the SR experiments and update the code to fix them. Please feel free to contact us if you have further questions.

[zwq1111]

Thank you very much for your reply! I have another question, I noticed that you have processed the dataset used accordingly when running the code. I am currently looking into this part and I was wondering if it would be possible to tell me how the dataset was processed.

[zhiyuns]

As described by the paper, when performing cross-modality synthesis, we first co-registrate the modalities using ANTsPy ( mode) and then prepare the h5 file using ./data/prepare_h5.py. If you want to perform super-resolution, please resample the HR image to get LR image, and upsample the LR image using to match the shape with HR image. After that, prepare the h5 file using ./data/prepare_h5.py. We also upload an example as ./data/ixi_demo.h5 for a clearer demonstration of this process.

[zwq1111]

Thanks again for your reply! I have one more question about the code. I have downloaded the required ixi dataset on the website, but I didn't find the 'T1_HR' and 'T2_FLAIR_linear' modality data in ./data/prepare_h5.py. I'd like to ask you for details about it.

[zhiyuns]

It's just an example, you can modify it to fit the actual task and situation. For IXI, you can modify it with ['T1', 'T2', 'PD'] to reproduce our result.