Such digital image processing technology might seem like just a fun game, but it has serious applications, such as helping law enforcement find missing children years later.
This particular software project, funded by Google and Intel, has tapped into thousands of photos of children and adults via an Internet search to gain knowledge about how faces age (based on pixel arrangement) regardless of the lighting, expressions and poses seen in the original photograph. An algorithm makes sense of that information and applies it to a new subject's photo to determine how the person in the picture would age, figuring in changes such as nose growth and eye narrowing. The software runs on standard computing hardware.
"Aging photos of very young children from a single photo is considered the most difficult of all scenarios, so we wanted to focus specifically on this very challenging case," said Ira Kemelmacher-Shlizerman, a UW assistant professor of computer science and engineering, in a statement. "We took photos of children in completely unrestrained conditions and found that our method works remarkably well."
The researchers say the system is more sophisticated than face-aging smartphone apps, like AgingBooth, and works a lot faster than a forensic artist can in missing person cases.
The research team's paper on the new technique has been posted and will present its findings at the June IEEE Computer Vision and Pattern Recognition conference in Columbus, Ohio. Also, see a video showing face morphing below.
The current edition of the software takes factors such as gender and age into account, and future versions will likely include additional factors such as ethnicity, wrinkles and hair whitening.
