Find: Google’s “Project Tango” is a smartphone with Kinect-style computer vision




Google’s “Project Tango” is a smartphone with Kinect-style computer vision
// Ars Technica

Google is launching yet another crazy moonshot project. This one is a prototype called "Project Tango," which squeezes 3D computer vision technology—similar to that used in the Xbox Kinect—into a smartphone. The device is being cooked up by Google’s Advanced Technology and Projects (ATAP) group, which just moved over from Motorola. Johnny Lee, the Technical Program Lead at ATAP, described the project:

Project Tango strives to give mobile devices a human-like understanding of space and motion through advanced sensor fusion and computer vision, enabling new and enhanced types of user experiences – including 3D scanning, indoor navigation and immersive gaming.

The computer vision is enabled by a new co-processor from Movidius, called the "Myriad 1." The chip was designed from scratch to bring Kinect-style computer vision to smartphones, where size and power-draw are a huge challenge. In fact, the man quoted above, Johnny Lee, is a former Microsoft employee and worked on the Kinect technology before jumping to Google. Google's goal with Project Tango is to produce the hardware, ship the phone out to developers, and see what they come up with. TechCrunch, which was pre-briefed on the device, says Google is giving the device out to 200 developers, and signups for access start today. Developers that apply will have to pitch their ideas to Google. 

My favorite video.

The computer vision isn't meant to enable Leap Motion-style hand waving for input, but to let the phone know where it is in 3D space. The rear of the phone is packed with sensors that would allow the device to "scan" a room and build a 3D model of it, which apps could interact with. This sounds like Google is making an augmented reality platform that could really tell what is in a room, instead of crudely guessing the room geometry based on a 2D camera feed. 

Read 2 remaining paragraphs 



via Mobiles @ NCSU http://ift.tt/OgBaL8