EyeGage wants to look into the eyes of auto workers, drivers

LaVonda Brown was working on her doctorate in electrical engineering in Atlanta a few years ago when she made a discovery that's now being reviewed by the auto industry.

Her work at Georgia Tech involved developing a robotic educational agent — basically a robotic tutor to keep students engaged through the use of eye tracking. The program could monitor where the students were looking on a tablet during a math test, for instance.

She later went to Emory University and applied the technology to Alzheimer's research.

During her research, Brown said, she was living with someone with a drinking problem. "And I guess this is a career hazard, but I would analyze his eyes during the day, and then at night when they might be red and glossed over and slow to respond," Brown told Automotive News.

It occurred to her there were commercial needs for that technology.

"I assumed my work was going to be in the educational field," she said. "But I did some market research and saw that drug and alcohol screening was an important market."

Brown, 33, founded the company EyeGage to provide a screening process that determines when individuals — auto assembly workers, drivers, industrial workers, delivery personnel — are under the influence of a substance and a danger to others. The software startup is housed at Georgia Tech, as part of the university's business accelerator program, CREATE-X. She spoke with News Editor Lindsay Chappell. Here are edited excerpts.

Q: What's the premise of your venture?

A: The idea is to determine whether a person is impaired. And that's something we can detect by looking into their eyes.

Your eyes are so informative, and you can tell a lot from a person's eyes.

How does it work? How does it detect a problem?

We take a picture of your eyes and get characteristics, like the size of your pupil, the color of your sclera — the white part — to see whether red blood vessels are present. We look at the height of the eyelids. And then we also look at how the pupil responds, because it's supposed to respond at a certain speed. If it's slower or faster, that tells us something. And also any involuntary twitching.

You've been in discussions about using this in the auto industry. Where is the potential there?

The technology could be built into cars. As you're driving down the road, or even prior to driving, it can detect whether or not you're under the influence of drugs and alcohol. We can embed our software into the cameras that some vehicles already have. But it might take a while to do that.

We've learned that there's also interest from the manufacturing side, where they're building vehicles. You could have our software on a mobile device, for instance, on a forklift in an industrial setting. Every time a user gets on that forklift, they scan their ID, or whatever their protocol is, and do a quick eye scan with that. So you would know whether an employee is under the influence before you give them access to drive it.

I can picture screening a large factory work force and imagine there could be pushback from employees. Maybe even some concerns about false positives, where the scan says I'm intoxicated, but I just have a cold.

We're always collecting data. The more diverse data we get, the more accurate our algorithms are. We're currently raising a few million dollars to bring all of this to reality. We have an event coming up where we will bring in people and give them alcohol. We'll test with a Breathalyzer, draw blood and get their readings, and then transport them back home once they've gotten back down to a low point of blood alcohol concentration.

We just had an event at a Georgia Tech football game, and we set up a tent to collect data from the people there.

For every person we test, we want to have at least 100 eyes to compare it to. When I take a picture of my eyes, I want to be able to compare them to 100 African American women in their early 30s to see whether mine are normal. But we also need eyes in the data of sober eyes, of people on cold medicine or people who have consumed a lot of caffeine.