I’m using Gentoo Linux. I have installed Shotwell 0.32.2 with support for OpenCV.
In Shotwell, I have opened a few photos, added a few faces using the button “Faces” at the bottom. And then trained a few faces via the meny “Faces / Train Face “” From Photo”.
Nothing happens after that. I would assume that it would find the person in at least some other photos or something?
Apologies, that is not very intuitive and needs a UX overhaul at some point. The Flow is as follows:
In the “Faces” button, mark a face, either manually or by clicking “Detect Faces…”
Name it, save it
Then use “Train Face from Photo”
The next time you click “Detect Faces” on a photo, it will try to use the trained faces to automatically name detected faces → You still need to click “Detect faces” on those fotos you want to tag faces in
The recognition rate / detection rate will improve if you use the DNN-based detection, which can be enabled as described in Faces in photos, last section
Thanks! I tried detecting faces, but it finds “faces” in basically anything from toilet paper to flowers. It could detect like 10 “faces” in a photo with only one person.