Currently detecting faces is a separate step which can become tedious since this is usually only done once per image / album and the albums must be selected manually (Digikam does not remember which images/albums it has already processed to skip them).

So the easiest way to process all images for faces *once* is to do it directly during import (in the background, after import has been finished). 

Would it be possible to integrate the faces detection dialog into the import workflow?