Bug 478557

Summary:	Face Recognition Feature Suggestion
Product:	[Applications] digikam	Reporter:	davidsrothman
Component:	Faces-Recognition	Assignee:	Digikam Developers <digikam-bugs-null>
Status:	CONFIRMED ---
Severity:	normal	CC:	cal030, caulier.gilles, michael_miller, thomasdn
Priority:	NOR
Version First Reported In:	8.3.0
Target Milestone:	---
Platform:	Microsoft Windows
OS:	Microsoft Windows
Latest Commit:		Version Fixed/Implemented In:
Sentry Crash Report:

Description davidsrothman 2023-12-15 14:33:43 UTC

1. The Algo must provide a score (or perhaps a set of scores) when it goes thru face recognition.  When the results are presented to the user, it would be very helpful to allow us to sort by that score or scores.  That way when 1000+ results are displayed, we don't need to hunt for the good/obvious matches (or similarly, reject the blurry unrecognizable ones).  That would be an incredible time save to users who are trying to process a lot of data.

2. When processing a lot of the face recognition results, a slip of the mouse can and does mark pics incorrectly - in some cases a selection of many pics.  Correcting via ^Z doesn't seem to work (or if it does, it isn't obvious to me).  I'm especially concerned about the false positives.  I don’t know if the Digikam saves date/time for each recorded recognition, but if so, sorting on that would allow the user to correct mistakes much more easily.  Alternatively,   If a suggestion is rejected, it returns when u search again.  If u ignore it, I believe that solves that problem, but if u already have a large ignore list (say u have a lot of group photos with a lot of people u want to ignore forever), then adding 'ignores' which u may not want to ignore forever is problematic.  Perhaps (and easy for me to say as I'm not coding it 😊) adding user-defined rejected bins (or maybe more practically a secondary short term reject bin might be an easy option)?

Great Product.  thanks for your consideration.

Comment 1 Thomas Damgaard 2023-12-15 20:34:54 UTC

This would be most welcome features! +1

Is it still possible to vote on bugs/features? I cannot seem to find the vote function ?

Comment 2 caulier.gilles 2024-10-08 09:32:16 UTC

Hi Michael,

The recognition score value must be already available somewhere in background with DNN model. Right?

Best

Gilles

Comment 3 Michael Miller 2024-10-08 10:35:15 UTC

Hi al,
The recognition score is computed one face vs another.  There isn't a global score for all faces, or even for a new face against a set of known faces.  This is so digiKam can recognize faces at all angles and profiles.  For example, if a known face is looking straight at the camera digikam can now recognize the same face turned more to the side and match it.  Once that face suggestion is confirmed, digiKam will use the new sideways face and match it to a face in full profile. The same process would repeat for faces turned the opposite direction.

The new facial recognition model coming in digiKam 8.5.0 is much more accurate at matching faces, even very blurry ones.

For item 2, I agree.  We need to provide a function to check faces within an approved set to see if there were any accidental matches.  Because the new model is much more accurate, even 1 accidentally approved incorrect face can generate many false-positive matches.

Cheers,
Mike

Comment 4 caulier.gilles 2024-12-01 09:04:01 UTC

digiKam 8.5.0 is out with many improvements in face detection and recognition. Please update these entry accordingly with this version. Thanks in advance...

https://www.digikam.org/news/2024-11-16-8.5.0_release_announcement/

Comment 5 caulier.gilles 2025-04-11 18:13:40 UTC

Hi,

The 8.7.0 pre-release Windows installer from today have been rebuilt from
scratch with Qt 6.8.3, KDE 6.12, OpenCV 4.11 + CUDA support, Exiv2 0.28.5, ExifTool 13.27, ffmpeg 7, all image codecs updated to last version (jxl, avif, heif, aom, etc.).

Please try with this version to see if your problem still reproducible...

https://files.kde.org/digikam/

Thanks in advance
Best regards

Gilles Caulier

Comment 6 Maik Qualmann 2025-07-25 05:47:17 UTC

*** Bug 507466 has been marked as a duplicate of this bug. ***