Government-funded researchers in Taiwan have developed a vending machine that recommends purchases based on people's faces, one of the inventors said on Friday.
The machine, designed by the Institute for Information Industry in Taipei, builds a profile after checking characteristics such as complexion and hair color, said researcher Tsai Chi-hang. Those clues help the machine guess a shopper's gender, approximate age and other things that might be helpful in promoting a suitable product.
Researchers spent the past year using a grant from Taiwan's Ministry of Economic Affairs to build the first machine, which was rigged up on Thursday to spit out free cosmetics samples in the institute's lobby.
The machine looks for clues like whether a person has glasses, a beard or a mustache, said Tsai. Based on that it guesses their use of make-up or frequency of shaving, Tsai said. It then might recommend a facial mask, razor, or health products that people in a certain category are statistically likely to buy.
"If you stand in front of it, the machine has ways of recognizing your characteristics, though it doesn't know exactly who you are as that would infringe on personal privacy," Tsai said in an interview.
"It's a new concept, so it's taking some time to catch on," he said, referring to a model machine installed this week in the institute's lobby.
Researchers in Japan unveiled a similar concept in August last year. The Taiwanese machine isn't a copy of that but the Taiwan researchers kept up on what Japan was doing, Tsai said.
The machine also attempts to detect any smartphones, e-readers or tablets the buyer might be carrying, the institute told Taiwan's semi-official Central News Agency. That recognition would tell the machine whether the shopper was equipped to download books, music or films.
Taiwan's institute aims to tailor-design machines for vendors, with storage capacity and exact features depending on the individual order, Tsai said. The institute has already been in touch with some vendors, he said, but he declined to say how much machines would cost.
Information on what buyers actually choose will be stored and sent to the Internet, helping retailers to analyze shopping patterns. The institute has previously researched WiMax technology and worked with China to develop the Android-based Ophone OS.