27 May 2015 02:30 PM EST
-by Yuliya Geikhman, Staff Writer; Image: Your iris is unique like your fingerprints. (Image Source: Michael Gil / msvg on Flickr / CC BY 2.0)
Real forensic work is rarely as easy as shows like CSI make it seem—there’s no way to enhance grainy surveillance footage until it’s crisp and clear, and no way to match a face to a name from just a photo of an eye in a rear view mirror… at least, until now.
New technology developed at Carnegie Mellon University uses iris pattern recognition to ID people by their eyes. The photo recognition technology can make the ID from a photo of an eye taken from up to 40 feet away—ideal for many cases in law enforcement.
Iris patterns are as unique as fingerprints, and iris scanners are becoming less of a sci-fi gadget and a more commonly used biometric. Carnegie Mellon’s scanner can differentiate between patterns and match them up to a database, even in poor lighting conditions or from a mirror reflection.
This eases the current quality constraints of iris scanners—the image has to be clear enough and taken from the right angle to be useful. Carnegie Mellon’s tech can also detect when something is in the way of the full iris image, be it an eyelid, eyelash, or anything else.
The team behind the technology has recently demonstrated in a video a possible use for the scanner—a photo of a partial face reflected in a side-view mirror, like the kind that would be taken at a traffic stop. The scanner can work from a still image or a video feed, making it applicable in a wide range of scenarios where a positive and quick ID is needed.
Marios Savvides, the team’s lead, points out that “there’s a lot of potential applications for saving lives, since you can identify a possible criminal wanted for murder or other crimes.”
Maybe we’re not that far from CSI-level tech after all.
-by Alex Compton, Contributing Writer; Image: Indiana cornerback A’Shon Riggins returns interception against Ohio State. (Image Source: IUHoosiers.com) Bloomington, IN—The Spring Game...