[digiKam-users] Influence of "ignore" and "no face" on face recognition

[Eduard Zalar]

Hi all,
I use digiKam for years and the new face recognition is really awesome! You
did a really good job there! It has detected more than 4000 faces just in
my pictures of 2020 and 2021.
And this is really overwhelming... Committing a face to a person tag is the
easy option. But I do not clearly understand what are the consequences of
the two negative options:

What happens if I select "No face"? Does this mean that the face detection
is influenced?
I can imagine using this option to suppress the detection of very unsharp
faces or faces which are mainly covered. Right now, such kinds of faces are
also detected - which is really awesome, but not useful from my point of
view. Am I right with this assumption and is the face detection training
data adapted by this?

What happens if I select "Ignore face"? From my understanding, this face is
not tried to be "recognized" anymore, even if I repeat a face recognition
run. But still, the detection will find this face again and will categorize
it again as "unknown". Is this assumption correct?

Final question for now: How often do you recommend "Clear and rebuild all
training data" with the maintenance tool? For example after detecting and
assigning x pictures?

Again, the functionality is really awesome! I just want to find a way to
adjust it for my needs.
Best Regards
Eddie
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/digikam-users/attachments/20210407/17706520/attachment.htm>