You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Just to make sure we don't duplicate work here: The optional mask would be the label image. And the association should be possible via the metadata of the ROI table & the obs dataframe of AnnData.
See the proposed OME-Zarr table spec: ome/ngff#64
My understanding is that we'd use the region_key in .zattrs to specify which column in the obs dataframe contains the relative path to the label image. For our use case here and the current implementation, each row of the dataframe would point to the same label image (but one could eventually also store e.g. a combined ROI table, where they point to the different label images in the different wells)
(branching from #45)
This requires #194
The text was updated successfully, but these errors were encountered: