[digiKam-users] digikam and facerecognition

[KRODER Gerhard]

ah, i see, grouping inside not recognized faces by "similar faces". A 
kind of similarity search (which exists already for whole pictures), but 
only on found faces. Basic tools for this could be avaliable already, 
but adopting migth be some effort. Maybe developers can tell...

Yes, I'd imagine this feature to be a benefit, and vote for this, too.

Cheers, Gerhard.


Am 09.02.2022 um 11:18 schrieb daj omu:
> If you ever have tried picasa, faces are recognised and categorized 
> into albums. See image enclosed. DK just puts every face into a big 
> unsorted album.
> I would like to see the same approach as for Picasa.
> image.png
>
> On Wed, Feb 9, 2022 at 11:10 AM KRODER Gerhard <gerdst1 at gmail.com> wrote:
>
>     Hi dajomu
>
>     i can't realy answer in depth, as i am a simple user too. But i do
>     not
>     quite understand your intention. What kind of grouping of faces do
>     you
>     expect/wish? Colour, brightness, environment, glasses, beard, hair
>     color? I do not think this is possible on the short hand. This
>     seems to
>     extend the scope of "face" recognition.
>
>     AFAIK faces are grouped already, but only as for recognized faces,
>     which
>     in fact helps a lot.
>
>     Cheers, Gerhard.
>
>     Am 09.02.2022 um 09:21 schrieb daj omu:
>     > Hi,
>     >
>     > anyone know if and when DK will group unknown/unconfirmed faces
>     > instead of showing them all mixed up? It does a pretty good job at
>     > detecting faces, but how it presents new faces makes it almost
>     > impossible to use.
>     >
>     > dajomu
>
>
-- 
Mit freundlichen Grüßen
Gerhard Kroder
--
Mob.: +49 160 92 45 32 61
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/digikam-users/attachments/20220209/25dd7cf5/attachment-0001.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image.png
Type: image/png
Size: 60398 bytes
Desc: not available
URL: <http://mail.kde.org/pipermail/digikam-users/attachments/20220209/25dd7cf5/attachment-0001.png>