Digikam 7.2 faces updates

[Rob D]

I just wanted to say I just tried the most recent 7.2 merge, and I like the "feel" of how the new faces interface works.  Great job students!

I do have a question regarding YOLO inclusion.  It is planned in the 7.2 release?  And is enabling it just a matter of changing:

            m_dnnDetectorBackend = new OpenCVDNNFaceDetector(DetectorNNModel::SSDMOBILENET);

to

            m_dnnDetectorBackend = new OpenCVDNNFaceDetector(DetectorNNModel::YOLO);

in facedetector.cpp (and installing the necessary support files)?

YOLO detects so many more faces, it would make me sad not to have the option.

Rob

[nghiaduong]

Hi Rob,

Above what you have specified, you also need to activate CMake script in https://invent.kde.org/graphics/digikam/-/blob/master/core/data/facesengine/CMakeLists.txt#29 , in order to copy YOLO data files to your system.

Regards,

Nghia

[Gilles Caulier-4]

Hi Rob,

I suggest reporting this wish to bugzilla, where we can talk about
safety as mailing threads can be difficult to follow in time.

There are some points to clarify as the huge Yolo data model file to
install with digiKam. This can bloat the target system for some users
and this can be problematic.

Thanks in advance

Gilles Caulier


Le jeu. 3 sept. 2020 à 20:30, Nghia Duong
<[hidden email]> a écrit :
>
> Hi Rob,
>
> Above what you have specified, you also need to activate CMake script in https://invent.kde.org/graphics/digikam/-/blob/master/core/data/facesengine/CMakeLists.txt#29 , in order to copy YOLO data files to your system.
>
> Regards,
> Nghia