Google to Make Smartphones That Recognize Faces

Soon, smartphones will not only take snapshots and videos but also recognize the objects pictured. Google is working with a Silicon Valley chip designer to let mobile devices do that kind of heavy computing internally rather than relying on remote data centers.

 

Alphabet Inc.’s best-known unit and semiconductor startup Movidius on Wednesday announced a collaboration aimed at bringing technology described as deep learning to handsets.

 

The search giant has already demonstrated image-recognition capabilities with Google Photos, a mobile app and image hosting service that stores and analyzes images uploaded from smartphones. Users can search through the photos by typing in the names of objects, like flowers or houses or mountains, or use a photo of a face to find others that contain it. But uploading images takes time, and searching through them from a phone depends on a wireless connection that may not always be available.

 

With built-in image recognition, smartphones could identify objects in real time for a variety of applications, like identifying people to authorize transactions, aiding blind people and translating signs.

 

Movidius, which specializes in computer vision, has previously worked with Google in an effort known as Project Tango that help create a kind of three-dimensional map of indoor spaces. Under the latest arrangement, Google will purchase an unspecified number of a Movidius chip called the MA2450 along with software to help exploit it, the companies said. Financial terms weren’t disclosed, nor did Google comment on how and when such technology might reach the market.

 

“By working with Movidius, we’re able to expand this technology beyond the data center and out into the real world, giving people the benefits of machine intelligence on their personal devices,” said Blaise Agϋera y Arcas, head of Google’s machine intelligence group in Seattle, in prepared remarks.

 

Read More: Google to Make Smartphones That Recognize Faces – Digits – WSJ