What Will Apple Do With Swedish Eye-Tracking Technology?
andylim writes "An article on recombu.com explores the possibility that Apple is gearing up to launch eye-tracking technology soon. Citing a patent filed in 2008 that mentions 'gaze vectors' and a recent purchase of units from a Swedish eye-tracking company, the author suggests that the inclusion of eye-tracking tech in the company's forthcoming tablet would be Jobs's magnum opus. 'What better flourish to a career that began with the popularization of windows, icons, mouse and pointer than to usurp them all?'"
There are two main problems with eye tracking. First, your eyes are always moving. Second, they’re attracted to motion.
Eye tracking, done correctly, would have to avoid both of these pitfalls. It would be possible, but tricky. It would have to differentiate between the constant motion of your eyes and deliberate motions that you wanted to make, or at least not be adversely affected by all of their unconscious movements. It would also need to avoid causing movement or changes on the screen that would draw your attention away from what you intended to look at.
For instance, if a normal cursor was displayed at the detected position of your gaze, it would (A) obscure, (B) distract, and (C) float irritatingly away from your gaze if its positioning was even slightly miscalibrated.
Alexander Peter Kristopeit bought his basement from his mommy for one dollar.
Full disclosure: I am a vision researcher who has worked with most of the available eyetracking systems on the market.
I had not heard of the company, and a quick look at their product line tells me why not: their standalone systems are limited to 60hz/120hz depending on the model - this was good several years ago, but has been considerably eclipsed by other companies' designs (e.g. S-R Research's Eyelink hardware, which happily does 2000hz monocular tracking). It looks on par with ASL's Eye-trac gear, which has similar limitations.
I would bet that Apple just bought a few of their systems to use in internal testing - I sincerely doubt that anyone there is using such slow gear for major research.