[digiKam-users] Organizing Photos Located By Facial Recognition

Jonathan Kamens jikamens at gmail.com
Fri Jul 29 15:01:55 BST 2022

This is a gross thing for MS to do.

Someone ought to stick it to MS for doing proprietary garbage like this 
by writing a tool which reads the contents of the SQLite database and 
writes the face tags into photos if they aren't already there.

On 7/28/22 5:26 PM, José Oliver-Didier wrote:
> The drawback with using the current version of Microsoft Photos is 
> that it does not read/write the face region metadata back to the file 
> as Windows Photo Gallery did (using the MS People XMP Schema) and 
> Picasa (Experimental MWG Face Region XMP Schema). Digikam supports 
> reading/writing both (MS People/MWG Face Regions).
> Microsoft Photos face detection/recognition is pretty good, the 
> drawback again is all the effort you put into grouping the faces is 
> kept in a local sqlite database.
> - J Oliver
> On Mon, Jul 25, 2022 at 5:13 PM Marc Palaus <marcpalaus at hotmail.com> 
> wrote:
>     I think digikam cannot group faces that you have not previously
>     identified. You have to first manually identify a few faces for
>     each person and then run the face detection again.
>     In the past I used Picasa for the first face recognition (has the
>     feature you ask for), but beware that it does not detect existing
>     keywords/faces, so these would be overwritten, potentially losing
>     information.
>     On Jul 25, 2022 16:43, Tony Schwartz <InelUnit at outlook.com> wrote:
>         Microsoft Photos does an ok job at grouping faces together
>         that it perceives are the same.  It just lacks the flexibility
>         and other options DIGI has.
>         Right now I am sorting through a 2 TB hard drive and having
>         Digikam pull out all the faces.   I bet at the end I will have
>         over 2,000 photos.  Out of that 900 will be ones I care about.
>         Is there a way to ask digicam to sort the photos like
>         Microsoft photo does?  Or with Digi I must simply tag them?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/digikam-users/attachments/20220729/50e10fbb/attachment.htm>

More information about the Digikam-users mailing list