Google's image-recognition AI just got way smarter.
The company announced a new computer-vision platform, Google Lens, which could give our smartphone cameras even more power.
CEO Sundar Pichai showed off Google Lens with an innocuous image of a little girl playing baseball, taken from behind a chainlink backstop. It was a cute pic — but the criss-crossing metal of the fence in the foreground was a bit of a bummer.
That is, until Pichai activated the Lens image and erased the fence from the foreground, with no degradation of the image of the little girl at the plate.
It even has an AR function, as Pichai then demoed the camera scanning a street full of storefronts, with their Google listing popping up automatically once the AI recognizes the space.
One of the most useful use cases was when Pichai showed a video of a smartphone user holding up their camera to a Wi-Fi router and Google Lens to scan the network name and WAP password from the label, instantly connecting the phone to the network. Impressive.
Lens will be available as a feature using Google Assistant or Google Photos before it hits other Google software, according to Pichai.
Using Lens with Assistant will give your AI friend a new visual component, allowing it to use images to do things like set up events in your calendar. The Photos integration will give you the ability to identify famous paintings, building, and even make phone calls directly to businesses from photos after their subjects have been identified.