Preventing blurry (but correctly identified) faces from degrading future facial matches

[Thomas Seaford]

Is there a way to tell the facial recognition feature not to use a face in
its training data? There are quite a few fuzzy faces in my library that the
program correctly identifies as a face, and sometimes even correctly
matches to the right person, but then the quality of future matches for
those people falls dramatically due to the blurry image. Essentially, I'm
looking for a way to say "Yes, you correctly identified this fuzzy blob as
a face and the face is this person. But don't use this image to identify
future matches of this person."

I know the recommendation has been made in the forums to use non-facial
regions when identifying people manually to avoid screwing up the training
data. However, in this case, the program independently determined the face
region. There are many such instances; I'm not sure manually removing or
reclassifying the facial region will be efficient. For those I don't care
about, I ignore, but that isn't desirable for all of them. I'd like to be
able just to tell the facial recognition feature not to consider these
images.

*Thomas Seaford*
*704.761.8156 <704.761.8156>*

"*In the beginner's mind, there are many possibilities. In the expert's
mind, there are few.*" *Shunryu Suzuki*
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/digikam-users/attachments/20240603/719e3bbd/attachment.htm>