When Mark Clifton, CEO of Princeton Identity, wants to use his phone, he takes it out and just looks at it. A black-and-white image of his eyes shows up on the screen, and the screen unlocks immediately — no passwords or thumbprint needed.
“People give technology like that one second,” says David Ackerman, chief scientist of Princeton Identity, the company that created the phone’s iris recognition system. “Anything longer than that, and they won’t use it.”
The unlocking trick is made possible by a tiny unit located next to the phone’s front-facing camera. A flash of infrared light, a fraction of a second long, illuminates the user’s eyes, and an infrared camera takes images of the iris, comparing them to an image stored in the phone’s data. If they match, the phone unlocks. Much like a fingerprint, every person’s iris is unique, so it can serve as biometric identification.
Clifton’s demonstration happened to take place on a Samsung Galaxy Note 7, the first phone to include this feature. As it happens, that particular model of phone was recalled due to a serious problem with exploding and bursting into flames, but the iris recognition system that makes its security features work will undoubtedly show up in new Samsung models. In fact, iris recognition is becoming a common security feature on all kinds of consumer devices.
And one of the leading iris recognition companies is, for now, located on the SRI Sarnoff Corporation campus on Washington Road, the same campus where color television was invented.
The story of SRI’s latest spinoff company, Princeton Identity, began in the mid 1990s when a British-American Cambridge professor named John Daugman invented a way to encode pictures of irises so that a computer system could quickly compare two photos to see if they matched. Or perhaps it began in 1953, when scientist F.H. Adler proposed using photographs of the iris, instead of fingerprints, for identification. Or maybe it goes back to 1892, when French scientist Alphonse Bertillon, inventor of the mug shot, had the same idea.
While the idea behind iris recognition is nothing new, the technology to make it a reality came about only in 1994 with Daugman’s mathematical invention, and only now has it reached the point where it is both cheap and practical.
SRI’s history with iris recognition shows just how difficult it has been to bring the technology from textbooks to the marketplace. The company, already known for its work with infrared cameras and sensors, was a natural fit to pursue commercial applications of iris recognition. In the late 1990s it spun off a company called Sensare that set out to develop iris recognition for banking ATMs. That company went bust because the computing power to make it work was too expensive and complicated.
Kevin Richards, chief engineer for Princeton Identity, says the decreasing size and cost of computer processors and electronic components has finally allowed a functioning iris scanning camera that can fit into a tiny space inside the guts of a phone.
And that’s why in September, Princeton Identity, a 30-person unit of SRI, was spun off into its own company, funded by Samsung. “A lot of it is taking advantage of today’s evolving technologies that have pushed down the size and cost of sensors and processors,” Richards says. “What we’ve been doing for a few years is getting smaller and less expensive components so that you could do the iris recognition on a very small platform, and inexpensively.”
SRI International Sarnoff is a nonprofit lab, the successor of the RCA Sarnoff labs that were at the forefront of electronics research in the mid-20th century and that formed the core of the Route 1 corridor’s commercial research institutions. Thousands of people once worked on the campus of postwar, institutional-looking buildings, but that is now down to about 200. That number will shrink even further when Princeton Identity departs later this year for its own headquarters somewhere in the Route 1 corridor.
Ackerman and the rest of the team who made the phone-based iris recognition system had been working on various iris-related schemes for years, and the company had already demonstrated several applications for its technology. The most elaborate is a high speed scanner meant for airports or other places where many people need to be identified quickly. Previous iris recognition platforms, such as those installed at SRI’s own labs, required the person being identified to stare directly at a camera and hold still for several seconds. Despite its advantages, that system is slow and cumbersome. SRI created a system called “Iris on the Move” that can identify multiple people as they walk past cameras.
Several different types of iris scanners made by SRI are used in the real world, including a security system that controls access at Busch Stadium for the St. Louis Cardinals, and another at a Middle Eastern construction company where thousands of workers have to identify themselves and clock in every day. Now that it’s on its own, and with an infusion of capital from Samsung, Princeton Identity plans to further refine and market these applications.
Ackerman believes biometric identification is the future of security, since it has several important advantages over passwords. “From a user perspective, you don’t have to remember anything,” he says. “You always have your eyes.” The iris recognition is considered secure enough that in addition to unlocking the phone, the Galaxy Note also used it to store private files, and even for making credit card transactions.
Ackerman, who joined SRI in 2005, says the core of the technology is the uniqueness of the human iris. The iris, the colored part of the eye, is what expands and contracts to control how much light is admitted to the eye. The unique folds and markings of each iris are fully formed by age two and change no further over a lifetime. No two eyes are alike, not even on the same person, and any two irises can be compared with only a one in 1.4 trillion chance of a “false positive” match.
Iris recognition works by taking an infrared photograph of the iris, which goes through the eye’s colored pigment and can see the textured surface of the iris underneath. Once the computer has a picture of the iris’s unique landscape — it kind of looks cratered, like the surface of the moon — it’s time for the analytical tools to go to work. A program takes the round image of an iris and converts it into a digital rectangular image, which can be represented numerically. Using the mathematical technique invented by Daugman, the numbers are compared to see if they represent the same iris.
In this part of the process, Princeton Identity’s system is little different from that of other companies doing the same thing. It is the infrared flash where PI claims its competitive advantage. “What we bring to the game is a unique aspect in that we actually use an infrared flash,” Ackerman says. “You really need to put a lot of brilliant infrared light as an illuminator to see through the pigment, but for a very short period of time.”
As anyone who has looked at a bright light knows, ultra-bright light can be damaging to the eye. Infrared light is a good heater, so extremely bright IR light could even cause an eye to boil like an egg in a microwave. The key to avoiding this hazard is the ultra-short time pulses, turning on and off again in rapid succession. So it is Princeton Identity’s infrared strobe light that has given it a leg up in the iris recognition game.
Image analysis is also key to the ability to make a system that actually works under a variety of circumstances. If someone is squinting their eyes or looking in the wrong direction, identification is much more difficult, but a computer program could still correct the skewed image and make a match.
Ackerman, a physicist, is a specialist in optics and a veteran of Bell Labs, where he worked for 22 years. Lasers that he developed there are still being used in telecommunications today. He grew up in Illinois, where his parents were art teachers. He has an undergraduate physics degree from Cornell and a doctorate from the University of Illinois. “I’m one of those split brained people who do art on one side and science on the other,” he says. Even on his days off, he can’t quite get away from optics: his hobby is photography, and he has won awards for his nature photos.
In science fiction, iris recognition technology is often depicted as almost magical in its ability to scan a subject’s eyes, even at a glance and even from a great distance, giving surveillance states an almost omniscient ability to track the movements of citizens. (U.S. 1 is guilty of this comparison, having cited the movie “Minority Report” in a previous article about SRI, February 4, 2015.)
Ackerman says this public perception of the technology leads to potential clients having unrealistic expectations about what it can and cannot do. “People look at ‘Minority Report’ and believe that iris recognition is possible from 10 meters and from any angle,” he says. In reality, the laws of physics limit what can be accomplished. The current record for distance is about 100 feet, and that was done using an eight-inch telescope. Any farther than that and the distortion in the air due to heat makes a photo of sufficient detail physically impossible.
And although “Minority Report” iris-scanning spider robots are not in the cards any time soon, Princeton Identity is exploring some unique applications for its hardware. In one of the company’s labs there is a prototype device that looks a bit like a police radar gun with the handle of a cordless drill. It’s a handheld scanner meant to be used to identify subjects who usually don’t carry good photo ID — horses. A few years ago a group of investors had funded SRI to build the scanner and sell it to horse racing tracks, to prevent sneaky substitutions.
The horse racing market didn’t pan out, but Princeton Identity is now investigating selling the system to veterinarians who need to identify particular cows or horses out of a herd of hundreds. Similarly, the Bill and Melinda Gates Foundation has proposed using iris recognition to identify people in crisis medical situations, such as refugee camps or war zones, where people may be incapacitated or can’t communicate due to a language barrier.
The launch of the Galaxy Note 7 was an exciting moment for the employees of Princeton Identity, despite the eventual recall. It was the first time the technology found its way into the hands of millions of people. “It has been a great milestone to be part of something that exciting,” Richards says. “This will provide an opportunity for something like iris biometrics to become part of everyday life.”
Princeton Identity, 201 Washington Road, Box 5300, Princeton 08543. 609-256-6994. Mark Clifton, CEO. www.princetonidentity.com.