The camera acts as the device's 'eyes', and the CPU acts as the 'brain'.
Providing deep learning solutions and computer vision from devices to the cloud

Written by Josh Walden, Senior Vice President and General Manager of Intel's New Technology Group


An era has arrived where devices that are not smart or connected cannot survive. With the emergence of connected devices that understand their surroundings and react accordingly, new solutions unlike those of the past are also appearing one after another.

Intel has been able to secure a leading position in computer vision and cognitive computing by building core technologies based on RealSense™. Simply put, computer vision can be understood as a technology that enables machines to visually process and understand their surrounding environment. The camera acts as the device's 'eyes,' the central processing unit acts as the 'brain,' and the vision processor functions as the 'visual cortex' that processes visual information. Integrating all of these enables the implementation of features that are attracting significant interest in various emerging fields, such as navigation, mapping, collision avoidance, tracking, object recognition, and survey analysis.

The introduction of RealSense-based depth-sensing cameras has brought about innovative technology that enables devices to 'see' the world in three dimensions. To support this paradigm shift, Intel has acquired companies in the fields of machine learning, deep learning, and cognitive computing, aiming to open up a world of amazing possibilities previously unimaginable, such as recognizing objects, understanding scenes, and performing authentication, tracking, and navigation. As devices become increasingly smarter and decentralized, and with the number of connected devices predicted to reach 50 billion by 2020, specific System on a Chip (SoC) attributes are expected to become even more critical to realizing human-like visual capabilities.

It is in this context that we would like to announce Intel’s ongoing acquisition of Movidius. By acquiring Movidius, Intel has secured a low-power, high-performance SoC platform, enabling it to accelerate the application of computer vision. Furthermore, through this acquisition, Intel will acquire algorithms suitable for deep learning, depth processing, navigation, mapping, and natural interaction, and will accumulate extensive knowledge in embedded computer vision and machine intelligence. Movidius technology will contribute to achieving results by enhancing and optimizing RealSense capabilities.

In particular, it appears that Movidius will play a significant role in the process of Intel implementing new technologies. With the ability to track, navigate, map, and recognize scenes and objects based on Movidius's low-power, high-performance SoC, new opportunities have been gained in fields where heat, battery life, and form factor are critical factors. Specifically, this technology will be applied to areas such as augmented, virtual, and mixed reality (AR/VR/MR), drones, robotics, and digital security cameras. Movidius's industry-leading computer vision SoC family is expected to complement not only Intel's existing diverse IPs and product roadmaps but also RealSense.

Computer vision will completely overturn the landscape of next-generation computing, and Intel will lead this change in computing based on its RealSense and cognitive computing technology product families through the acquisition of Movidius.