海角大神

Some will crow and others will lament how their crow鈥檚 feet prematurely age them after uploading a photo of themselves to Microsoft鈥檚 age-guessing聽. But this fun website presages a future in which computers can make sophisticated predictions and decisions based on who they're looking at.

Felipe Rose, who has appeared in Native American garb as one of the disco stars聽聽since the 1970s, is 61, but was delighted to find out聽 that Microsoft's algorithm sees him as age 33.

鈥淥h it鈥檚 amazing ... I feel like I鈥檓 30. I can party and dance like a 20-year-old, and living life without any drama helps your blood pressure and stress lines,鈥� Mr. Rose said in an interview after getting his results. 鈥淢ore importantly, never forget to live, love, and laugh.鈥�

The hashtag #HowOldRobot is now trending on Twitter.

However, this is more than just a party trick.

Microsoft's site is using huge amounts of source data to determine the apparent age of a face, a field known as machine learning,聽according to Dr. Karl Ricanek, facial recognition software researcher and developer at the聽聽at the University of North Carolina in Wilmington, N.C.

This type of machine learning relies on a massive number of source images to create a model of what aging looks like on human faces. Using algorithms to mine the troves of data users upload to the internet daily is an increasingly important and competitive field for tech giants like Microsoft,聽,听补苍诲听, according to Dr. Ricanek.

鈥淭hey [Microsoft and others] want to make these very simple and automatic tools to manage your 'face life' [the ways in which facial recognition can be used in technology],鈥� Dr. Ricanek said. 鈥淲e are a culture of selfie-takers. We document every minute of our lives with pictures and videos and whoever can come up with a system to organize and be able to search through this big data is going to win large. They鈥檙e going to draw all the traffic to their tools. And they鈥檙e going to make a lot of money.鈥�

Dr. Ricanek, whose work has been used by the FBI, says that a similar site,聽 was based on his work at University of North Carolina at Wilmington. Face My Age takes an uploaded photo plus information about age, gender, and ethnicity to arrive at a judgment about a user's "face age," how quickly or slowly a face is aging.

鈥淚f you鈥檙e 30 and a smoker, you might be surprised to find that your face age is 42,鈥� Dr. Ricanek told the Monitor. 鈥淪moking dries the face, causing premature wrinkles, aging it.鈥�

鈥淲ith my software what we鈥檙e doing is telling you how old your face looks relative to others your age,鈥� he added.

According to a聽聽Corom Thompson and Santosh Balasubramanian, the engineers said they "wanted to create an experience that was intelligent and fun could capture the attention of people globally."

Dr. Ricanek concludes that while the How Old Robot and other apps like it may be a fun way to needle friends about their looks, ultimately the goal of such advances is to create recognition technology that would allow devices to more intelligently interact with users. For example, a TV that restricts what can be watched based on the age of the viewers in the room. 聽

鈥淭here鈥檚 going to be a lot of integration of this technology in our everyday lives,鈥� he said. 鈥淲hether it鈥檚 to unlock our telephones or confine the channels my child can see because it knows 鈥極h, there鈥檚 a young person there now,鈥� or it refuses to play an 18-or-older video game because Xbox says 鈥極h, this person is 10 and this game is for someone 18. There will be vending machines that recognize you and know your preferences, and ATMs that recognize you.鈥�

At some point, he added, 鈥渨e鈥檙e just going to become accustomed to them. We won鈥檛 have to say 鈥榖iometrics鈥� or 鈥榬ecognition鈥� because it will just be part of everyday life.鈥�