Trying out #YuNet #FacialRecognition on #DigiKam for the first time tonight. We have a lot of faces in our photo library. Apparently YuNet is more accurate but slower than #SSD, well I guess I am about to find out. I do love the #snap of digikam, so I get the latest without any effort..
Some exciting developments with #facialdetection for my #shysculpture . I got the #yunet #neuralnetwork working and wow- it's so effective compared to the cascaded classifiers. It even works with masks. Now it's set up to dim the neon - to hide- when it sees a face.
This all works locally- no faces or live video feeds sent to big corporate servers. It's main flaw is a two second delay- perhaps this will be fixed with a compute stick or faster processor.