Intel Launches RealSense ID for Facial Recognition on Device

Intel today launched the newest addition to RealSense, its range of depth products and tracking technologies designed to give machines depth perception capabilities. Called RealSense ID, it is an in-device solution that combines an active depth sensor with a machine learning model to perform facial authentication.

Intel claims that RealSense ID adapts to users as physical characteristics such as facial hair and glasses change over time and works in various lighting conditions for people “with a wide range of heights or complexions”.

But several studies and analyzes by VentureBeat itself of public reference data have shown that facial recognition algorithms are susceptible to various biases. One problem is that the data sets used to train the algorithms distort white and male. IBM found that 81% of people in the three collections of face images most widely cited in academic studies have fairer skin. Academics have found that technology and photographic techniques can also favor lighter skin, including everything from sepia films to low-contrast digital cameras. As a result, Amazon, IBM, Microsoft and others have imposed moratoriums on the sale of facial recognition systems.

In response to this and other criticisms, Intel says RealSense ID, which starts at $ 99 and will launch in the first quarter of 2021, has built-in anti-spoofing technology to protect against false entry attempts using photographs, videos or masks . The company also says it has a false “one in 1 million” acceptance rate and processes facial images locally, encrypting user data, and is activated through “user awareness”. In other words, it will not be authenticated unless requested by a pre-registered user.

Intel sees RealSense ID initially being used for smart locks, access control, point of sale, ATMs and kiosks. The company plans to expand into subcategories such as finance, health and smart access control.

“We did extensive data collection from all ethnicities in Asia, Europe, the Middle East and Africa,” Intel’s head of product management and marketing, Joel Hagberg, told reporters during a briefing this week. “We were very careful to ensure that all ethnicities were covered.”

Corporate Vice President Sagi Ben Moshe added: “We invested a lot of money to build a proprietary database … It was a massive data capture project that took us a long time and that is what makes it unique”. He and Hagberg did not disclose the size of the database or whether audits were carried out to identify biases it may have, nor the method by which the photos and subjects contained in the database were selected.

RealSense ID follows the launch of the Amazon Web Services Panorama Appliance, an Amazon plug-in device that connects to a network and identifies video streams from existing cameras. The Panorama Appliance comes with computer vision models for manufacturing, retail, construction and other industries, with support for models built on SageMaker. It also integrates with AWS IoT services, including SiteWise, to send data for further analysis. Along with the Panorama Appliance is the AWS Panorama SDK, which allows hardware vendors to build new cameras that perform computer vision at the edge. It works with chips designed for computer vision and deep learning by Nvidia and Ambarella, and Panorama-compatible cameras will readily work with AWS machine learning services.