Julien Cohen-Adad
Julien Cohen-Adad
> I just updated the config.json file used by dcm2bids so that it uses the ImageComments field contained by the json sidecars to identify the phase and magnitude niftis. it...
After discussing internally, it seems like going with a specific CLI per method will be easier for maintenance/implementation in the long term, arguments: - shorter CLI files - less potential...
> We have to check if we want to live with the fact that Windows is not supported. we can definitely live with it, especially because WSL (and now WSL2)...
indeed tsv is part of bids convention— might be overkill to require that. or we could generate it with our dcm2bids wrapper
> Does magnitude greater than 2 exist in the bids convention? hum... good question. This is quite uncommon (most ppl use 2 echoes). Is there some bids-hack we can do...
Possibly relevant to the discussion: - testing data download for parallel testing: https://github.com/neuropoly/spinalcordtoolbox/issues/2957 - about caching (+ integrity testing if data already present): https://github.com/neuropoly/spinalcordtoolbox/issues/2959
Amazing! Thank you for this very useful feature @hermancollin
> Perhaps it would be best to add "orientation" to the MODELS dict for each of the nnunet/monai models, then specify LPI/RPI/AIL on a per-model basis, then fetch this value...
> The bad news is, the problem is deeply embedded into SCT, because Image is a core part of SCT, and it pre-loads the data everywhere. We're probably seeing this...
Good question. Given that the code already creates one label per lesion, I _think_ the approach would be to run the statistics per label, and for each label, have the...