Google, Movidius Team to Give Devices 'Human Vision'

A smartphone camera can see and capture. Google and Movidius want to add what's missing: context, intelligence, and a sense of space and motion.

Google and Movidius, a designer of ultra-low-power vision processor chips, designs, software and more, are working together to create "intelligent, vision-based applications" that will significantly push the capabilities of today's smartphones, the pair announced Feb. 20.

As part of the agreement, Movidius' Myriad 1 vision processor platform will power Google's Project Tango—an Android-based smartphone product and development kit created by Google's Advanced Technology and Projects (ATAP) group that strives to give machines a human-like understanding of space and motion.

Myriad 1 is Movidius' first-generation Vision Processor Platform, a "new ultra-low-power, high-performance and programmable architecture of computational chips, software and development tools that enabled a range of devices to intelligently understand and contextualize their surroundings," Movidius said in a statement.

Compared with traditional processors, Movidius said, it can add 10 times the flexibility, speed and power efficiencies.

The Motorola ATAP group made headlines last October when it introduced its Project Ara modular phone design. When Google agreed to sell Motorola to Lenovo, in January, it decided to extract ATAP from Motorola and reabsorb it within Google. (According to Movidius CEO Remi El-Ouazzane, Google CEO and co-founder Larry Page developed the ATAP group before Google's acquisition of Motorola.)

El-Ouazzane, speaking with eWEEK, offered a deeper explanation of Movidius. The company's ultimate goal, he said, is to give devices "human vision."

"When you look at things, you are capturing, tracking things in motion and you are taking millions of 3D measurements. What a camera in a mobile device can do today is only the capture piece. What we're talking about is adding all of that extra intelligence," said El-Ouazzane.

"What we want to do is change your camera from dumb to intelligent, so it can not only capture [what it sees] but also extract information," he continued. "But to do this intelligence extraction, you are talking about a hundreds-of-gigaflops operation. It is the most intensive thing you can think of in computing."

Movidius is interested in enhancing a range of technologies, from autonomous cars to wearables.

"Our vision—no pun intended—is to enable human vision into the devices we are invested in, and to put all of that compute power into a small footprint."

The Movidius platform, he explained, has three areas of development. One is intelligent vision—putting locational vision and context understanding into devices. The second is power efficiency, and putting everything possible into the smallest possible footprint. ("It's everything for us.") The third is cutting-edge apps.

Project Tango, said El-Ouazzane, will represent a paradigm shift, making possible the "next wave of mobile apps on the market."

He offered examples from assisting the visually impaired (an app that uses the camera to "see" and offer audio guidance) to virtual gaming apps that can see a player's living room and create a window into an immersive, virtual reality within the room.

"These are the types of apps that Google will be talking about at launch, but they are the tip of the iceberg," said El-Ouazzane.

Movidius doesn't have an exclusive relationship with Google and is always exploring new areas, El-Ouazzane said. "We have been working with [Google] because they are by far the ones working in a very advanced way."

"I am personally super-excited about robots, about [artificial intelligence], about wearables," he said. "The fact that I'm seeing my technology adopted for mobile technology is wonderful ... but for me, the way it can be used goes way beyond smartphones."

Follow Michelle Maisto on Twitter.