-
Notifications
You must be signed in to change notification settings - Fork 119
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
Hypothalamus Implementation #459
Conversation
I was getting an attribute error with the current import of the yacs library |
That looks like a typing bug. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I went through the code and have some comments.
Some of the points can probably be done by someone else, but there are a couple of questions we want to fix soon, before you invest much more work. Specifically, the questions regarding bias field correction.
please don't merge from dev, do a rebase instead. Rebase - of course - should not be done if we are officially collaborating on a branch it is checked out by other people. but if this is mainly on your fork, just rebase, which will add the dev changes before your changes and keeps the history linear. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This review was "overdue". Now it is down to me to fix run_fastsurfer.sh.
b572bae
to
b168655
Compare
cec07c7
to
7b576d5
Compare
I think the branch is ready to merge. Or something else needs to be tested? |
I think there is one issue left that I need to resolve, which is based around conforming/pre-processing of the T2 image. In the current code, the conversion to uchar is independent of the bias field correction. This is not necessary, the bias field script has a The next/other open issue is the processing if the I have a branch where both of these issues are started (actually 1 should be finished, 2 is started but not tested yet). Finally, there used to be a sphinx documentation problem, that I was supposed to fix, but did not get around to yet. So both are not really critical issues, but more "side-issues". |
I think it would be nice to merge my Branch to dev before OHBM ( Monday, June 24), as I will be presenting HypVINN. In the preproc.py, before I registered T1 and T2, I ensured that the T2 image was converted to uchar and that robust rescaling was applied. I don't do this if they put none as a registration step, so I agree that improving the T2 conform will be nice. Let me if I can help with something |
We can also merge now and do the rest as separate PRs? |
without __pycache__ files include HypVINN labels Create Main function for N4 bias correct include fs bin for registration
Include copyright statements in Hypvinn files
update run pipeline test fs
formatting
formatting
Add hypvinn url file
- rename mode "multi" to "t1t2" - reformatting for line length - replace single quotes to double quotes - add typing information - clean up docstrings - replace os.path with pathlib.Path
Fix typo in N4_bias_correct.py
- remove/rename changed filenames - format table HypVINN/config/checkpoint_paths.yaml - add config HypVINN/data_loader/data_utils.py - fix typing and formatting HypVINN/utils/checkpoint.py - fix YAML_DEFAULT HypVINN/utils/mode_config.py - set default values for get_hypvinn_mode HypVINN/inference.py - fix inclusion of ModalityMode HypVINN/run_prediction.py - move HypVINN/run_pipeline.py into run_prediction.py - fix typing, e.g. FileBasedHeader - fix function parameters - add help text to hypo_segfile argument - fix passing of t1_path and t2_path - various other changes
- Fix errors in t1t2 mode - Fix qc_snaps argument in run_fastsurfer.sh - Fix reg_mode test in run_fastsurfer.sh
update model weight (version 1.1) for hypvinn update download script
Convert the T2 image to unit8 to save hard disc space for registration - set --uchar for N4_bias_correct.py - `conform.py --no_force_(img_size|vox_size|lia) --t2 <arg>` if --no_biasfield Copy the T2 file to $SUBJECTS_DIR/$SID/mri/orig/T2.001.mgz
- Added docstring for functions in HypVINN
- adding space before colon separating parameter name and type
…of a constant Make saving a main thread function instead of a future/task
- Fix for bash-3.2 (|& => 2&>1 |) - Add warning if T2 is passed, but biasfield is skipped --no_biasfield
- add comment paper vs. retrained weights. - add hypothal module in main README - fix sphinx errors - add an additional message in fix_links to debug errors - add the hypvinn script documentation
I am going to merge this now. This should be ready for the base use case, but I want to fix the robust scaling of the T2 image. This can be a different PR though. |
New
TODO