It’s fun going through baby pictures and seeing what you used to look like back in the golden days. But what could be even more entertaining is the exact opposite — knowing what you’ll look like 50 years from now — and now researchers from the University of Washington have found a way to do just that.
By analyzing thousands of photos of people’s faces on the Internet, this software can predict what someone will look like decades from today based on only one photo.
While there are existing technologies out there that do something similar, the UW researchers say their method is the first approach that can process visual aging data even with photos that have different types of light, poses, and facial expressions.
“Our extensive user studies demonstrated age progression results that are so convincing that people can’t distinguish them from reality,” Steven Seitz, a UW professor of computer science and engineering, told UW News. “When shown images of an age-progressed child photo and a photo of the same person as an adult, people are unable to reliably identify which one is the real photo.”
Other than the fun aspect of this neat program, researchers say their software can speed up the process of rendering older versions of missing children and provide more accurate images, “for which few prior results have been demonstrated.”
The research team, which was funded by Google and Intel, just published a paper and will present its findings in June at the IEEE Computer Vision and Pattern Recognition conference in Columbus, Ohio.
Ira Kemelmacher-Shlizerman, a UW assistant professor of computer science and engineering, told us that the public will soon be able to test out the technology via webcam and photo upload. We’ll update this post when that happens.
