Persona is one of the latest fashion magazines in Tokyo. It's printed on heavy stock paper and is full of photos of models and clothing. The only thing missing is text.
You "read" Persona with a smartphone camera.
An app recognizes the images, queries a cloud database and then downloads related information such as pricing and availability of dresses. Other photos feature images of tomatoes and wine, triggering a related vegetable delivery service and wine retailer, as well as online coupons.
NEC's Gaziru image-recognition software is being used to link the cloud database to both the online edition of the quarterly magazine and its print version, which is to be distributed to 10,000 beauty salons across Japan next year. The app and the magazine launched earlier this month, with 18 companies participating.
The concept is similar to other image-recognition and augmented reality apps such as Google Goggles, but NEC says its system is more focused on affiliated products and services instead of Internet search results. It's particularly speedy when searching a database of images that have been linked to objects beforehand.
"The response time is really quick," said Ken Tanoue, an NEC manager involved with the project. "You could also use this to scan a program or tickets for a concert with this and download additional information."
Tanoue pointed his smartphone with a Gaziru app at a TV showing a video of models walking down a catwalk. As the video showed a different model, the app recognized each one and instantly displayed her name.
The video was processed by NEC before the demo, but Gaziru, first developed about two years ago, can also recognize unregistered objects such as a shirt on display in a shop or food such as a dish of pasta.
In those cases, it will query the cloud database and retrieve information about things that resemble the object in question. For instance, it will recognize a dish of meat-sauce pasta and call up a recipe for spaghetti.
The technology is rooted in the same know-how that NEC has used to develop its NeoFace face-recognition system, which can be used as biometric security for accessing PCs.