CyberLink’s facial recognition technology will soon be appearing in more IoT devices thanks to a new partnership with ASUS. The latter will be integrating CyberLink’s FaceMe engine into its own Tinker Board 2 single-board computer (SBC), which supports a wide range of IoT and AIoT applications.
The ASUS Tinker Board 2 utilizes 64-bit Armv8 technology, and is built with a six-core Rockchip RK3399 system-on-a-chip (SoC) and a multi-core Mali-T860. The SBC runs on the Android 10 operating system, and offers wireless connectivity in addition to dual 4K screen output capabilities.
FaceMe, meanwhile, is a highly accurate facial recognition engine that has consistently performed well in independent NIST testing. The engine is capable of both face detection and face extraction, and can identify registered faces (or profile unregistered ones) in real time.
The new partnership will enable facial recognition in devices that run on the Tinker Board 2. The two companies believe that the joint solution is particularly well suited to the retail and hospitality industries, where it can be deployed in kiosks, point-of-sale terminals, and access control systems. It can also be installed in digital signage to track demographic data like age, gender and emotion, all of which can be analyzed to improve customer-facing operations.
The Face Recognition Edge AI Dev Kit comes with FaceMe and the SBC, giving developers everything they need to add facial recognition to their own devices. The project is a collaboration between CyberLink and ASUS’s IoT team.
“Facial recognition is one of the most popular technologies powering an exponentially growing set of IoT solutions in a booming market,” said Dr. Jau Huang, CEO of CyberLink. “We are honored to partner with ASUS and provide solutions developers with one of the most accurate and flexible facial recognition SDK in the world.”
CyberLink’s FaceMe engine has already been integrated into Vypin’s access control kiosks, and the Nx Witness video management solution from Network Optix. The facial recognition solution is still able to identify people who are wearing masks.
(Originally posted on FindBiometrics)