By snapping up the startup, Intel will look to bring its perceptual computing platform to virtual reality headsets, drones, robots and driverless cars
Intel has announced it will buy Movidius, a startup specialising in computer vision processors commonly found in smartphones, virtual reality headsets and drones.
The acquisition is effectively an extension of Intel’s RealSense strategy, which has seen the company create a technology platform for powering computer vision systems and sensors to enable devices, such as driverless cars, to make sense of their surroundings through a technique called Perceptual Computing.
“Simply put, computer vision enables machines to visually process and understand their surroundings. Cameras serve as the “eyes” of the device, the central processing unit is the “brain,” and a vision processor is the ‘visual cortex’,” explained Josh Walden, senior vice president and general manager of Intel’s New Technology Group.
“Upon integration, computer vision enables navigation and mapping, collision avoidance, tracking, object recognition, inspection analytics and more – capabilities that are extremely compelling in emerging markets.”
Buying Movidius will give Intel access to company’s system-on-a-chip (SoC) platforms that are specifically designed to power computer vision applications, as opposed to Intel’s own processors that are more commonly associated with a large range of computing tasks.
For example, Movidius’ latest SoC, the Myriad 2, fits a high-performance, low-power consumption processor with a programmable architecture onto a fingernail-sized chip. This means the SoC can be used in devices where access to power is contained, such as small drones, yet still allow computer vision systems to run. The SoC is notably used with Google’s Project Tango to power augmented reality (AR) software, currently found in Lenovo’s Phab 2 Pro smartphone.
With access to these chipsets, Intel’s purchase of Movidius is likely to give it the scope to push RealSense into more devices, from virtual reality (VR) headsets in the consumer world to smart security cameras for industrial sectors.
Movidius also specialises in algorithms tuned for deep learning, depth processing, navigation and mapping, and natural interaction, which Intel will be able to use to bolster the capabilities of RealSense.
Intel is not shy about its ambitions to build out RealSense as the perceptual computing platform of choice for companies looking to create smart devices.
This year has seen the chipmaker showcase Project Alloy, a reference platform for a wireless VR headset that packs all the sensors and chips needed to power it within the headset, rather than relying on external sensors or a PC to provide the processing grunt.
The headset comes with built-in sensors and 3D cameras powered by RealSense to track the wearer’s motions and allows for VR to be blended with AR to create a ‘merged reality’, a mix of real world objects in virtual environments.
Project Euclid is another RealSense based device, only this time using the platform to bring perceptual computing to robots through a chocolate-bar sized device self-contained PC that manages to pack in an Intel Atom processor, a RealSense camera, motion sensors and a communications array. Effectively, Project Eucilid gives robot makers access to the technology they need to integrate computer vision systems into their machines without the need to build bespoke hardware.
While Eucilid is due for release early 2017, with the Movidius acquisition more projects and reference platforms from Intel are likely to emerge and will likely bolster its efforts in the Internet of Things arena as well.
How much do you know about future tech? Take our quiz to find out!