Hey there. I know this might be several years late at this point, but I had a question that someone here may be able to answer.

Back when I had my N9, one of the features I loved was the face recognition in the photos application.

Well, I'm wondering more and more about that. How was it achieved? Did it use an external API? Was it internal?

And more importantly, can it be extracted and made into an android or jolla app?

Does anyone have any more information on it? Just curious in general.

It's highly doubtful you'll get the source for the face recognition from within MeeGo, as "open" as folks proclaimed it was, the apps were not subject to being open whatsoever in most cases.

A quick look around, I found face tracking and recognition online, a lot in GitHub which could be forked.

Tracking and detection
Object (and face) detection
Face recognition API

And many, many others. Good luck on your search though for the MeeGo sources - who knows, perhaps a disgruntled Nokia engineer might supply code to you.

But it seems as if webcam (with permissions) and JavaScript face tracking and detection has come a very long way since I last looked at it in 2012. Heck, I'm actually impressed with some of the features now available.

object detection for blinds
http://aipoly.com/

Be careful: for android & IOS

squirrel recognition http://www.i-programmer.info/news/10...on-and-ai.html

I demand an upgrade to c... c... c... (gosh it is so hard to make myself write that hated word)... CATS!

Hm. Interesting problem, as -- well, let's call them "c-beasts" -- come in variety of colors laid in various patterns, effectively wearing disruptive camouflage. One application of image recognition to such beasts I am aware of was based on shape -- instead of normal image, a shadow was analysed (http://www.quantumpicture.com/Flo_Co...lo_control.htm) -- and the shapes are quite similar across the population. Of course, there's famous Google c-recognition algorithm, but I bet the big G has some more power -- both raw computational and engineering -- that our small community.

On the other hand, let's consider goals, not means... If you are simply trying to keep your yard c-free, I heard that application of tiger dung in the perimeter works wonders. Try to obtain some and please report back the results

And the real challenge, I think, would be dog recognition, including not only established breeds, but mutts as well.

Thank you, briest. So instead of c-dung I will have the garden stinking of t-dung. Delightful.

Besides, I have seen an experiment on TV with a dog and lion urine. It did not work. The dog came, sniffed the line, and crossed it without a second thought. How could the poor beast know what creature made it if it had never seen a lion?

For clarity, you're writing about recognition of the face's owner's identity, right? Not just that a face is present? Or did I misunderstand?

I thought he was talking about face recognition in the photos. I always turned it off since it made the device run considerably slower.

Yes, face recognition in the photos. I never really noticed a change in the device speed, but it was a long time ago regardless.

Thanks for the info though, this is some great stuff.

I'm not really a developer, but it might push me to consider something.