`CorrectedImageStack` is insufficient for modern motion correction

[rly]

https://github.com/NeurodataWithoutBorders/nwb-schema/blob/19c4bbb36effb1454253072f70ebf2c83f2ea095/core/nwb.ophys.yaml#L335-L360

• xy translation is required but modern motion correction is often not a simple xy translation (e.g., NoRMCorre)
• if a particular software package was used, we should store metadata about the software used.
• users may not want to store the original image stack (just the corrected images, not the original or translation)
• users may not want to store the corrected images (just original + translation metadata)

Given the three different modes above for storing such data (corrected only, original + translation, all), would it make sense to have different types for these? Otherwise, all the fields have to be made optional.

cc @bendichter @oruebel @CodyCBakerPhD

[rly]

Partial duplicate of https://github.com/NeurodataWithoutBorders/nwb-schema/issues/228

[bendichter]

Given the three different modes above for storing such data (corrected only, original + translation, all), would it make sense to have different types for these? Otherwise, all the fields have to be made optional.

What did you have in mind?