This is such shitty reporting by TechCrunch. 1. The face detection APIs were new in 5.0 – although under NDA, they are hardly newsworthy, especially considering now that 5.0 has been in beta for weeks if not months. 2. Face detection is NOT the same as face recognition. There's an important distinction: The latter recognizes a particular face where as the former can only recognize facial features and detect that it is a face. 3. 9to5 published this story on Monday. It took TC 3 days to notice and rip it off? Lame.

Hopefully this will improve on the algorithms they have in Lion (which seemed to work ever-so slightly better than OpenCV's face detection and work poorly on profile views).

I'm wondering if this will trickle into unlock screens and profile logins, I can only see that being highly frustrating. I have little faith in biometrics and can just imagine that not being able to log in due to having your angry face on can only become a viscous circle of lockout.

Then again, if you can teach it your intoxicated face then you could have a phone version of gmail's mail goggles.

On a more serious note, can anyone give me a practical non entertainment use for facial recognition in an desktop app? With the quality of camera it's likely to be working with in the near future there doesn't seem to be much scope for more than just 'yep, that's a face, here's roughly where the features are'.

I keep wondering when the "Welcome back, Mr. Yamamoto" scene from _Minority_Report_ will become reality.

iOS including strong face recognition will be a big step toward that.

Not face recognition, face detection. The new APIs [1] make it possible to detect if an image contains faces and returns position data of various face features (eyes and mouth). There is no way to recognize which person a face belongs to.

As an aside, why do rumor sites always have to make it seem as if these things were somehow hidden? It's a well-known and documented part of a public API (it just happens to be under NDA at the moment, but it's not any kind of secret, really).

[1] https://developer.apple.com/library/prerelease/ios/#document...

Kinect can do that now.

Must be depressing to work at Apple and be the dev on the Photo Booth app.

Yes, I'm going to be that guy, but Google showed off face "detection" (not recognition) at Google IO, and in my opinion, in a fairly cool way.

Sorry that I actually read the source material (or is it merely that I implied that it's not some god send from Apple). The only conclusion that is supported by the literature, source tweets, or "revealed" API functions... indicate "facial detection". Not recognition as commenters here or the TC article imply. Maybe someone else has info they're not sharing with me.

Downvote away, feel free to actually indicate where this is new from information that was known or anything that supports the "facial [recognition]" that we're all speculating about. What a joke.

It's nothing new. Face detection, recognition and real-time tracking was demoed a year ago on the N900, so it's certainly possible: http://personalpages.manchester.ac.uk/staff/philip.tresadern...

It’s an article about a new feature in iOS. It’s factually wrong and awful but it doesn’t claim anywhere that the face detection is revolutionary.

What was your goal by bringing Android into this? That can only result in unproductive discussions and it’s also kinda irrelevant.

Because the reaction to this on HN is like most other things Apple announces after others already have, and as often is the case (and not helped by the article) it misattributes magical features that aren't even in the feature being discussed. I guess I should've just kept my groaning to myself.

You might be a victim of confirmation bias. Not everything should be seen from the Google vs. Apple angle.