New demo for globalnet

[YipengHu]

Subject of the feature

@jvhoutte suggested to add a demo using the global net predicting affine transformation. The alterntive may be to add an option and trained network in one of the existing demos.

What are you trying to do and how would you want to do it differently?

Still open to suggestions, but simply adding an trained affine network option.

Is this related to an issue/problem? To a bug report #533

Has the feature been requested before? If yes, please provide a link to the issue. no

If the feature request is approved, would you be willing to submit a PR? (Help can be provided if you need assistance submitting a PR)

yes

[mathpluscode]

Which dataset are we considering here? @jvhoutte @YipengHu

[YipengHu]

Suggest:

• a new set of results for the paired lung ct demo:
• edit code: https://github.com/DeepRegNet/DeepReg/tree/main/demos/paired_ct_lung
• update doc: https://github.com/DeepRegNet/DeepReg/blob/main/docs/source/demo/paired_ct_lung.rst

According to the config here: https://deepreg.readthedocs.io/en/latest/docs/configuration.html#globalnet Not sure it needs to edit any functions at this stage, but the related functions may be here: https://github.com/DeepRegNet/DeepReg/blob/main/deepreg/model/backbone/global_net.py

[YipengHu]

@acasamitjana As a student is interested in looking at #703, are you interested in looking at this ticket instead?

[acasamitjana]

Okey. In your previous comment, are you suggesting to update the paired_ct_lung demo or to create a new one using the same dataset?

[YipengHu]

Okey. In your previous comment, are you suggesting to update the paired_ct_lung demo or to create a new one using the same dataset?

either is fine, maybe start with changing the paired ct? if we identify anything better, we can always update

[acasamitjana]

extract_levels parameter is deprecated. How should we deal with that? Maybe opening an issue targeting some RC milestone? I believe that applies to GlobalNet as well as UNet, right?

[YipengHu]

extract_levels parameter is deprecated. How should we deal with that? Maybe opening an issue targeting some RC milestone? I believe that applies to GlobalNet as well as UNet, right?

GlobalNet should have never used extract_levels. You can safely ignore this for GlobalNet.

[stale[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.