CyberLink partners with Advantech for drastically improved edge facial biometrics performance
CyberLink founder and CEO Dr. Jau Huang spoke about the enhancements in the biometric performance of the company’s FaceMe technology running at the edge in artificial intelligence of things (AIoT) systems at the recent 2020 Advantech Embedded-IoT World Partner Conference.
The performance enhancements are the result of a successful strategic partnership with Advantech, according to the announcement, with FaceMe running on Advantech’s AIR Inference System with an Intel Movidius VPU. FaceMe’s deep learning algorithms deliver real-time facial recognition with a True Acceptance Rate (TAR) of 99.5 percent at 0.0001 percent False Acceptance Rate (FAR).
Huang spoke during the “TechTalks: The Evolution of Edge Computing and AI” segment of the conference, saying the partnership’s results show the leadership of both companies in the global AIoT industry. Cyberlink also touts its results in the NIST FRVT as showing its algorithms among the most precise and fastest facial recognition engines.
In a demonstration of the Advantech and Intel system running FaceMe, FaceMe’s performance was 20 times better than in a similar system with a conventional CPU, CyberLink says. Edge-based FaceMe saves bandwidth and computation costs compared to cloud solutions, in addition to the potential performance benefits.
“Leading hardware developers and system integrators around the world strive to innovate and develop new AI solutions. We welcome collaboration initiatives that integrate FaceMe’s AI facial recognition engine and bring reliable, secure and highly performing offerings to the market,” said Dr. Jau Huang, CyberLink’s founder and CEO. “CyberLink and Advantech’s partnership is strengthening both companies’ global AIoT leadership.”
CyberLink recently announced improved financial results so far for fiscal 2019, and its results in an AI challenge for the beauty industry.
Related Posts
Article Topics
artificial intelligence | biometrics | biometrics at the edge | CyberLink | edge computing | facial recognition | IoT | stocks
Comments