Bug 460439

Summary:	Keep tag, but delete face information
Product:	[Applications] digikam	Reporter:	Jens <jens-bugs.kde.org>
Component:	Faces-Workflow	Assignee:	Digikam Developers <digikam-bugs-null>
Status:	REPORTED ---
Severity:	wishlist	CC:	caulier.gilles, metzpinguin
Priority:	NOR
Version First Reported In:	7.9.0
Target Milestone:	---
Platform:	Appimage
OS:	Linux
Latest Commit:		Version Fixed/Implemented In:
Sentry Crash Report:

Description Jens 2022-10-14 16:58:25 UTC

SUMMARY:
Sometimes face detection finds a view of a person where the face is not very well visible, but I'd like to keep the tag to be able to find this person again (e.g. when wearing a mask, or from behind, or upside down, or ...).
In the faces workflow I'd like an option to delete the face tag but keep this same tag on the photo. Maybe as a fourth button below the thumbnail.

Is this possible and feasible?

SOFTWARE/OS VERSIONS: KDE Neon 20.04 + Digikam Appimage 2022-10-02

Comment 1 Maik Qualmann 2022-10-14 18:02:29 UTC

The People view is based on faces (regions), not tags. If you delete the face region the image will disappear from the People view, but if the tag were left it would be only viewable in the Tag view.

Maik

Comment 2 Maik Qualmann 2022-10-14 18:07:43 UTC

The "Ignored" function is actually intended for faces that you don't want. You can still find the image under "Ignored", but the face region found is no longer considered by digiKam, nor is it written to the metadata (there is no standard in the metadata for ignored faces).

Maik

Comment 3 Jens 2022-10-14 22:28:08 UTC

OK, maybe this wasn't entirely clear.
I know about "Ignored" - this is for people I do NOT want to tag, at all.
But what about images where the face detection found "somebody", and I want to tag this image, but NOT use this face to train the recognition image - maybe because it's blurred, or half obscured, or the person wears a mask?
Currently I use non-face tags to do this, but it's suboptimal.

Reason:
I found that tagging only a few bad images can make face detection performance significantly worse.  Of course, in a perfect world, the image recognition engine should not be so easily confused, but being able to tag faces without Digikam using this face for training may also help.