CyberLink improves ranking in Nist testing
10 July 2019 14:03 GMT

Biometrics firm CyberLink announced on 9 July that its FaceMe AI Facial Recognition Engine has ranked among the top tier in the latest Face Recognition Vendor Test (FRVT) in the WILD 1E-4 benchmark category conducted by the U.S. National Institute of Standards and Technology (NIST).

With its algorithm achieving an accuracy rate of 97.02%, CyberLink’s FaceMe team ranked 12th among all test participants. The WILD 1E-4 test reflects use cases in Smart Retail environments, Smart Surveillance, and Smart City scenarios. The outstanding performance of FaceMe® further demonstrates its exceptional reliability and precision for implementation in a wide range of IoT/AIoT edge devices.

NIST’s FRVT serves as the world’s preeminent benchmarking authority for facial recognition algorithms, with participation from development teams around the world. The NIST FRVT WILD 1E-4 dataset consists of faces extracted from surveillance camera footage or photos, encompassing a wide array of real-world situations including a range of capture angles, poor lighting, or partially covered faces. The image variability simulates real-world use cases where systems would be required to accurately identify individuals in multiple different settings.

In the latest NIST report released in July 2019, FaceMe achieved a 97.02% accuracy rate (2.98% FNMR/False Non-Match Rate). Among all the world-class participants submitting NIST benchmarks, CyberLink’s FaceMe® team ranked 12th, only 0.27% away from the top-ranked team1, proving yet again that FaceMe® is a world-leading facial recognition engine.

“Since its release, FaceMe  has achieved outstanding results in major facial recognition competitions, including NIST and the MegaFace Challenge. The results confirm FaceMe®‘s global leadership in the facial recognition industry, providing world-class reliability, accuracy, and performance,” said Dr. Jau Huang, CEO of CyberLink. “The WILD 1E-4 test reflects real-world usage scenarios. The results of this test allow system integrators and developers to identify the most accurate and reliable facial recognition technology to implement in smart surveillance and smart retail scenarios.”

Industry Events