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
Heya, just curious as to why the return type here has changed. I think vigra (which is used internally) will always return uint32 and that way it was easy to "remain" in the vigra regime. Now one has to cast again. Are you working on replacing vigra for the watershed?
Cheers
D
The text was updated successfully, but these errors were encountered:
I have been using the output in some cases where uint32 might not be sufficient (i.e. ids > 4 billion), in order to avoid subtle issues I changed it here.
Hey thanks for the clarification. Do you think it would be possible to add a flag to this function to retain the original behavior? I'd like to avoid having to copy the data twice around.
Do you think it would be possible to add a flag to this function to retain the original behavior?
Sure. I see two options: either add a flag, or add an argument that specifies the output dtype (and that defaults to uint64 so that the current behavior is the standard).
elf/elf/segmentation/watershed.py
Line 164 in 9979a9d
Heya, just curious as to why the return type here has changed. I think vigra (which is used internally) will always return
uint32
and that way it was easy to "remain" in the vigra regime. Now one has to cast again. Are you working on replacing vigra for the watershed?Cheers
D
The text was updated successfully, but these errors were encountered: