Bug 446106

Summary: Visually identical pictures not detected as similar
Product: [Applications] digikam Reporter: Dor <dbenari2000>
Component: Database-SimilarityAssignee: Digikam Developers <digikam-bugs-null>
Status: RESOLVED WORKSFORME    
Severity: normal CC: caulier.gilles, metzpinguin
Priority: NOR    
Version First Reported In: 7.3.0   
Target Milestone: ---   
Platform: Microsoft Windows   
OS: Microsoft Windows   
Latest Commit: Version Fixed/Implemented In: 7.4.0
Sentry Crash Report:
Attachments: Picture 2 (high res)
picture 1
Picture 2 (high res)

Description Dor 2021-11-26 08:57:30 UTC
Created attachment 143958 [details]
Picture 2 (high res)

SUMMARY
***
NOTE: If you are reporting a crash, please try to attach a backtrace with debug symbols.
See https://community.kde.org/Guidelines_and_HOWTOs/Debugging/How_to_create_useful_crash_reports
***


STEPS TO REPRODUCE
1. Put the 2 pictures in same album
2. Find similar on one of the pictures (any percentage)

OBSERVED RESULT
The other pictures is not detected as similar

EXPECTED RESULT
The  other pictures detected as similar

SOFTWARE/OS VERSIONS
Windows 10


ADDITIONAL INFORMATION
Comment 1 Dor 2021-11-26 08:58:16 UTC
Created attachment 143959 [details]
picture 1
Comment 2 Dor 2021-11-26 08:59:41 UTC
Created attachment 143961 [details]
Picture 2 (high res)
Comment 3 Maik Qualmann 2021-11-26 11:28:28 UTC
With a similarity setting of 95% -100%, digiKam recognizes both images as identical without any problems. My guess is that you haven't updated the fingerprints, you have to do this every time new images are added or created.

Maik
Comment 4 caulier.gilles 2021-11-26 14:32:12 UTC
Hi,

Same s Maik here. I tried different similarity settings under Linux and images are detected with threshold over 95%.

Best

Gilles Caulier
Comment 5 Maik Qualmann 2021-11-28 16:40:14 UTC
Any feedback here? Fingerprints updated?

Maik
Comment 6 Maik Qualmann 2021-12-03 19:56:11 UTC
No feedback, it works for Gilles and me, I'll close the bug.
Note: In the meantime, the recognition of the reference image has been improved.

Maik