Eye-tracking handsets are one step closer to reality, courtesy of an improved iris recognition algorithm. The boffins over at the Dartmouth College managed to come up with an advanced iris recognition algorithm which has been christened as the Eyephone (cheeky dig at Apple, eh?). The algorithm currently runs on Nokia’s 810 tablet that is powered by Linux. Facial tracking algorithms are not new, but they tend to suffer on cellphones in terms of performance due to arm movements, constantly changing background light and composition. The Eyephone will be able to track the position of the eye relative to the screen as opposed to the direction of the person, applying an “error box” to track the user’s iris, achieving a 76%success rate in the process (in daylight), but driopping by 16% when the user is walking. More time will be required to make the algorithm all the more accurate, and be prepared to see more interactive apps on the cellphone in the near future if this catches on.
Filed in . Read more about Tracking.