Bug 392010

Summary:	Face recognition assigns faces to people without confirmation, leading to false positives
Product:	[Applications] digikam	Reporter:	MarcP <iwannaberich>
Component:	Faces-Recognition	Assignee:	Digikam Developers <digikam-bugs-null>
Status:	RESOLVED FIXED
Severity:	normal	CC:	caulier.gilles, iwannaberich, jmzd3an
Priority:	NOR
Version:	5.9.0
Target Milestone:	---
Platform:	Microsoft Windows
OS:	Microsoft Windows
Latest Commit:		Version Fixed In:	7.0.0
Sentry Crash Report:

Description MarcP 2018-03-18 19:23:10 UTC

The title is pretty self-explanatory. When performing face recognition, some faces are automatically assigned to existing people. This should be ok if face recognition had a 99.9% accuracy rate, but this is not the case. 

In practice, false positives are common, and there is no indication whatsoever that a face has been assigned to a person until you visit that person's faces by chance.

A simple confirmation before automatically assigning a face to a person would solve this issue. "e.g. "Does this face correspond to Mr. XXX?"

Comment 1 Jmz D3an 2018-08-24 14:41:14 UTC

I run into this problem often too. I suggest using the unconfirmed tag to solve this. There faces could be shown which are auto recognized and addigned but need user confirmation.

Comment 2 MarcP 2019-09-01 12:10:40 UTC

I believe this bug is no longer relevant (or won't be relevant soon). Faces are never assigned to people, just suggested. Therefore, metadata is never written for new faces without the intervention of the user. When I opened this bug, I was still not sure of that.

However, for the sake of clarity, I would not close it until Bug 392016 is solved (which will be, once the changes from gsoc2019 will be incorporated to the main branch).

Comment 3 caulier.gilles 2019-12-23 15:35:03 UTC

7.0.0-beta1 is out with new Face Recognition algorithm based on Deep
Learning/Neural Network API from OpenCV

https://download.kde.org/unstable/digikam/

Please test and give us a feedback

Thanks in advance
Gilles Caulier

Comment 4 caulier.gilles 2019-12-25 15:26:50 UTC

See the Neural Network example from digiKam 7.0.0-beta1 performing faces
recognition with 2 items face-tagged in first.

https://i.imgur.com/HHF4T9X.png

1 new item is recognized, name-tagged, highlighted with green frame, and linked
in "unconfirmed" virtual tag for confirmation.

Recognition work well using Deep Learning and give 97% of true positive
results. Look well my comments on the screenshot. I close this file now.

Merry Christmas and Happy new year.

Gilles Caulier