Bug 387271

Summary: Adding Objects and Scenes Recognition for BQM.
Product: [Applications] digikam Reporter: Yehia Amer <yehia2amer>
Component: Plugin-Bqm-AutoTagsAssignee: Digikam Developers <digikam-bugs-null>
Status: RESOLVED FIXED    
Severity: wishlist CC: caulier.gilles, neoderhacker, quochungtran1999, timokau
Priority: NOR    
Version First Reported In: 5.8.0   
Target Milestone: ---   
Platform: Other   
OS: All   
Latest Commit: Version Fixed/Implemented In: 8.3.0
Sentry Crash Report:

Description Yehia Amer 2017-11-24 16:46:57 UTC 
Digikam is Great Tool that and i think it is one of the first tools to support face recognition.
but Numbers of photos started to be really crazy, I will reach 500,000 photos soon and we need a new way to Automagically categorize these images not only by people & places but also by things

I think the next step now is to use AI & deep learning to:
1) detect things , Subjects , Scenes, and occasions  
2) add tags to photos & Albums (example: http://edenphotos.io/)
3) Search for photos based on who’s in them or what’s in them, like strawberries or sunsets
(examples: https://www.apple.com/lae/macos/photos/ & https://www.synology.com/en-global/beta/2017_Moments_1_0)
4) use these algorithms to make Face Recognition Even better

To do so we can use a lot of open source tools & CNN like: 
https://github.com/beniz/deepdetect/ , specially Inception v4 & ResNet
https://www.elastic.co/blog/categorizing-images-with-deep-learning-into-elasticsearch
https://github.com/Microsoft/CNTK/

Thanks a lot.
Comment 1 caulier.gilles 2023-12-01 04:26:58 UTC 
Hi,

With next digiKam 8.3.0 release, the auto-tags assignment feature have been
implemented without using a cloud service. The processing is done in
core application with delegate neural network models stored in computer.

For more details about auto-tags assignment feature, look on student work
report :

https://community.kde.org/GSoc/2023/StatusReports/QuocHungTran#Add_Automatic_Tags_Assignment_Tools_and_Improve_Face_Recognition_Engine_for_digiKam

Best regards
Gilles Caulier