If you knew when you were going to die, would it help you decide whether to sock more into your IRA or buy that new boat?
Karl Ricanek thinks it would, so much so that he is pursuing a market for people who want research-backed estimates for individual life expectancy.
A person’s face gives away those life expectancy clues, said Ricanek, a professor at University of North Carolina Wilmington and director of the ISIS Institute-Face Aging Group lab at the school.
Unhealthy habits over time like smoking and certain chronic conditions leave their markers on a face, and the right, detailed algorithms can help researchers determine lifespan estimates, Ricanek said.
Ricanek (at left
), who has studied facial recognition and age modeling in particular for more than a decade, teamed up with Jay Olshansky, an expert in the field of longevity.
“We’re combining my understanding of the face and the information it conveys, and his understanding of life tables and actuaries,” Ricanek said.
Ricanek and Olshansky, a professor at the University of Illinois at Chicago’s School of Public Health, are now co-owners and technical consultants of Lapetus Software Inc., a spin-off company that launched in June. UNCW has a small equity position in the new company, Ricanek said.
Last month, Lapetus Software released a website – FaceMyAge.com
– that the duo hopes is the first step in a suite of business products based on its research.
Ricanek points to a recent visit to his financial planner as an example of how personalized estimates could help with making decisions.
“The first question he asks me is ‘When do you plan on retiring?’ Probably 67,” Ricanek recalled. “And then the question is ‘How long do you think you’re going to live?’ I really don’t know.”
So, his planner input the assumption of 92 years old.
“There’s no data to support 92 as a general lifespan for anybody in the U.S.,” Ricanek said. “I can really use this on a daily basis to help people understand.”
This work in facial analytics extracts information by examining the face, though not necessarily to find the person’s identity as in facial recognition, Ricanek said.
He said Lapetus Software is looking to raise about $3 million in venture capital in two years, with the initial product likely targeting the financial services sector.
The new website, besides establishing a presence in the market, also helps the researchers improve their algorithms.
At this stage of the site, users can upload their photos – they have to be wearing minimal makeup and not smiling to help the analysis of facial points – answer some questions, and then receive a basic snapshot of the life expectancy for people like them.
Ricanek said the ability to age a person’s photo will be added to the site as the company gains funding.
The end use of this type of age progression and analytic software is wide open, Ricanek said, with potential customers ranging from insurance to cosmetics industries.
“There are many, many applications that we don’t even know about,” he said. “That’s why it’s important to construct and deliver the system in the most malleable way.”
But what about overcoming the creepiness factor, convincing people face their own estimated mortality?
“We’re born, and then we die. That’s just the plain fact of the matter,” Ricanek said.
The algorithm, he said, just helps figure out what to do in the middle part.