Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Implement recon models #123

Merged
merged 72 commits into from
Oct 18, 2021
Merged

Implement recon models #123

merged 72 commits into from
Oct 18, 2021

Conversation

georgeyiasemis
Copy link
Contributor

Reconstruction Models added:

  • Unet for reconstruction
  • EndToEndVarNet
  • KIKINet
  • LPDNet
  • XPDNet
  • JointICNet
  • AIRS model

Additional models added:

  • MWCNN
  • DIDN
  • NormUnet
  • Crossdomain
  • ConvGRU now is independent of RIM

@github-actions github-actions bot added documentation Improvements or additions to documentation python labels Oct 12, 2021
@jonasteuwen
Copy link
Contributor

This is addressing #116. Can you look into pylint?

Copy link
Contributor

@jonasteuwen jonasteuwen left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Some initial comments.

direct/data/transforms.py Outdated Show resolved Hide resolved
direct/data/transforms.py Show resolved Hide resolved
Comment on lines +792 to +793
assert_complex(coil_data, complex_last=True)
assert_complex(sensitivity_map, complex_last=True)
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Is this assert_complex still needed? Can't our internal representation completely switch to complex tensors, and use .is_complex() or so (that can be done in the assert_complex of course)?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This will require to change the whole software & models so let's keep this as a future change

direct/data/transforms.py Outdated Show resolved Hide resolved
Comment on lines -281 to -287
# 2D data is batched and contains keys:
# "filename_slice", "slice_no"
# "sampling_mask" of shape: (batch, 1, height, width, 1)
# "sensitivity_map" of shape: (batch, coil, height, width, complex=2)
# "target" of shape: (batch, height, width)
# "masked_kspace" of shape: (batch, coil, height, width, complex=2)

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It feels we should write this down somewhere. Maybe in the documentation?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Documentation of engine?

direct/nn/crossdomain/multicoil.py Outdated Show resolved Hide resolved
direct/nn/didn/didn.py Outdated Show resolved Hide resolved
direct/nn/didn/didn.py Outdated Show resolved Hide resolved
direct/nn/didn/didn.py Outdated Show resolved Hide resolved
direct/nn/jointicnet/jointicnet.py Outdated Show resolved Hide resolved
@jonasteuwen jonasteuwen self-requested a review October 18, 2021 15:44
Copy link
Contributor

@jonasteuwen jonasteuwen left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM. We need some small changes for v0.3, but good for now.

@jonasteuwen jonasteuwen merged commit 312bea6 into main Oct 18, 2021
@jonasteuwen jonasteuwen deleted the implement-recon-models branch October 18, 2021 15:48
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
documentation Improvements or additions to documentation python
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants