Seperate preprocessing from registration

[adamltyson]

Currently the preprocessing is fixed, and based upon serial2p datasets. The preprocessing should be a seperate step, potentially leveraging existing napari plugins.

Options could include:

• No preprocessing (load preprocessed data)
• Existing filters for serial2p datasets
• LSFM-specific filters - @horsto did you manage to get brainreg working on your lightsheet data?
• fMOST - @noisysky - it would be great to bring your fMOST preprocessing steps into brainreg/napari somehow so that others can use it on their data.

[noisysky]

@adamltyson Sure, I'll try to make it a bit more general (get rid of hardcoded values). Any guidance on how to incorporate it to brainreg-napari will be greatly appreciated. Should it be a command-line argument like --preprocessing fmost?

[horsto]

@adamltyson - indeed I got it working with our LSFM data, quite well actually. I did not add any preprocessing but rather fixed some of the imaging parameters which helped to clean up our data and make the alignment process smooth.

[adamltyson]

@noisysky - I think that's probably a good place to start, if we have the different preprocessing modules within brainreg, we can add the plugin functionality later. Perhaps this file could be refactored so that we have a collection of standalone functions, and then wrapper functions like preprocess_stpt, preprocess_fmost etc.

[adamltyson]

@horsto great to hear. I'd love to see the results sometime!

[adamltyson]

Closing in favour of https://github.com/brainglobe/brainglobe-registration