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
I think it would be useful to expose other methods of projecting along axes for ImageStack (e.g., DataArray.sum() and DataArray.mean()). This would add support the ExFISH case for projecting across (R, C) before spot segmentation (#1301).
Implementation
Some possibilities:
I think the thing that is the least work is to make a filter, Filter.Project that takes projection type as an input argument. This could replace Filter.MaxProject.
We could do the same as (1), but make it a method of ImageStack. I think a lot of the notebooks would have to be updated, as they use ImageStack.max_proj().
We could also make one filter component per projection type, but this seems too granular.
As a related aside: I think there should be only one way to max project (i.e., ImageStack.max_proj() and Filter.MaxProject are redundant).
Proposal
I think it would be useful to expose other methods of projecting along axes for
ImageStack
(e.g.,DataArray.sum()
andDataArray.mean()
). This would add support the ExFISH case for projecting across (R, C) before spot segmentation (#1301).Implementation
Some possibilities:
Filter.Project
that takes projection type as an input argument. This could replaceFilter.MaxProject
.ImageStack
. I think a lot of the notebooks would have to be updated, as they useImageStack.max_proj()
.As a related aside: I think there should be only one way to max project (i.e.,
ImageStack.max_proj()
andFilter.MaxProject
are redundant).Thoughts, @dganguli , @ambrosejcarr ?
The text was updated successfully, but these errors were encountered: