adamltyson
commented
5 months ago Now the code to transform cells from raw data to atlas space has been merged into brainglobe-utils, I think it's time to tackle the first "layer interpreter", assigning cells to brain regions within napari. This is a priority, because it's one of the few parts of BrainGlobe that's only available via a CLI (brainmapper).
This will require two things:
- Transformation of cells from raw data to atlas space
- Assign cells to brain regions and summarise, analyse etc.
This will essentially allow cellfinder and brainreg to be combined within napari, allowing analysis of cellular distributions within a whole brain image, entirely within a GUI.
This was previously attempted in cellfinder-napari (https://github.com/brainglobe/cellfinder-napari/pull/117, https://github.com/brainglobe/cellfinder-napari/pull/153). It shouldn't go there anymore, but there's a question about where it should now go (alongside other "layer interpreters). Originally I had thought that it should be a seperate package, and brainglobe-utils should have minimal dependencies. However, realistically, brainglobe-utils will always need to work alongside napari.
So @brainglobe/active-devs, are you happy to turn brainglobe-utils into a napari plugin, and begin to expose analysis functionality as napari widgets?
alessandrofelder
[From grant proposal] Napari provides a standardised framework for image analysis, as plugins operate upon “layers”, representations of data such as images and segmentation masks. We will develop tools to combine and convert these layers into information in a neuroanatomical context, such as summarising segmented cells as the number of cells per brain region. Developing these “layer interpreters'' to act upon fundamental napari data structures will allow current and future napari plugins to be used for neuroanatomical analyses.