Apologies if you have seen this before, but it blew my mind:

http://economictimes.indiatimes.com/tv/TED...ted/5231080.cms

It takes to about 6:30 to really kick off, but is worth watching from the start.

Holy shit.

I hadn't seen this video before, but I've seen the technology. I'm not sure why they prefer projecting onto physical objects rather than a more HUD-esque approach (either using goggles/glasses or direct retinal projection), which obviously gives less visual crosstalk when you have multiple users. The technology probably isn't quite there yet (or at least is too heavy for everyday use), but I don't see why it couldn't become as ubiquitous as Bluetooth headsets for receiving mobile phone calls are these days.

It seems like the closest thing we have to this technology right now are high-end mobile phones that have a high quality display, camera and enough CPU for it to all hang together - there are already a number of augmented reality apps for these but they almost all rely on location or orientation services (GPS/compasses) rather than the camera itself, which limits their use to tasks such as navigation, tourism and astronomy.

There are some limited image recognition-based augmented reality applications (although not many for mobile phones) but these tend to rely on putting easily-recognisable physical tokens on objects rather than recognising the objects themselves (which is obviously much harder). As image recognition gets better, I could imagine (for example) navigation applications that work on the underground (where clearly GPS doesn't work) by recognizing signs, etc.

Cheers,

Matt

It would be interesting to know how much of the demo was faked up or made simpler by requiring obvious tags to aid recognition.