Welcome to CyberLink Member Zone

Mar. 22, 2022
CyberLink's FaceMe® Achieves Perfect Score in iBeta's Advanced Level 2 Anti-Spoofing Test and Complies with ISO 30107-3

A 100% spoofing-prevention rate for both 3D printed and resin facial masks, confirms FaceMe® as a leading facial recognition solution for preventing biometric fraud in remote applications, such as online banking, requiring identity verification before granting access to sensitive data or valuable assets.

TAIPEI, TAIWAN — March 22, 2022—CyberLink Corp. (5203.TW), a pioneer in AI and facial recognition, announced today its FaceMe® facial recognition solution achieved a perfect score in iBeta’s Presentation Attack Detection (PAD) Level 2 test, indicating its compliance with the ISO 30107-3 testing and reporting requirements. This distinction requires protection from higher-level fraud attacks with 3D masks, and FaceMe® achieved a perfect score at 100% spoofing prevention, solidifying its position as a leading facial recognition tool for combatting biometric fraud.

iBeta is one of the world’s few agencies accredited by US National Institute of Standards and Technology National Voluntary Laboratory Accreditation Program (NIST NVLAP) for biometric testing. iBeta’s PAD test is an international benchmark for spoofing tests as it confirms to ISO-IEC 30107-3, an industrial standard for facial recognition technology’s anti-spoofing capability.

FaceMe® achieved iBeta’s PAD test level 1 compliance in late 2021, which focuses on presentations from 2D photos and videos. The level 2 test uses more sophisticated presentation attacks from 3D masks, such as 3D printed masks, resin masks, or latex masks. FaceMe® has achieved an Attack Presentation Classification Error Rate (APCER) of 0%, meaning the technology is extremely reliable in preventing attacks from any form of 3D masks. In addition, the Bona Fide Presentation Classification Error Rate (BPCER) is as low as 1.5% on iOS devices and 2.5% on Android device. In other words, FaceMe® achieved respective success rates of 98.5% and 97.5% when the test subject was a real person, placing it among the top facial recognition solutions for preventing biometric fraud attempts from a variety of techniques. Robust anti-spoofing is necessary to validate users’ identity for remote applications such as online banking, shopping for age-restricted products or even logging to the intranet at work.

"Biometric fraud prevention is one of the most critical imperatives for sectors of the economy such as fintech, where a security breach can take catastrophic proportions, but also wherever users are demanding remote access to sensitive data, fueling an explosive demand for remote authentication," said Dr. Jau Huang, CEO of CyberLink. "The iBeta PAD test is widely recognized as the global standard for testing the spoofing prevention performance of facial recognition technology offerings. Our perfect score at iBeta’s Level 2 test, which attests of its compliance with the ISO 30107-3 testing and reporting requirements, anchors FaceMe® as one of the very best facial recognition solutions to protect access from biometric fraud, even when presented with the most sophisticated attacks from 3D masks."

FaceMe® is optimized to run across hardware configurations, from high-end workstations to low-power chipsets frequently used in IoT and AIoT devices. It is the most versatile and adaptable facial recognition offering on the market today. FaceMe® provides solution builders and system integrators a fast, reliable, precise, and flexible facial recognition technology that can be deployed across multiple scenarios including security, access control, public safety, fintech, smart retail and home protection.


CyberLink FaceMe® iBeta Presentation Attack Detection Report:



For more information on FaceMe® SDK, please visit: https://www.cyberlink.com/faceme/sdk/overview

About CyberLink
Founded in 1996, CyberLink Corp. (5203.TW) is the world leader in AI facial recognition technology and multimedia software. CyberLink addresses the demands of consumer and commercial markets through a wide range of solutions, covering AI facial recognition, digital content creation, multimedia playback, video conferencing, live casting, and mobile applications. With years of research in the fields of artificial intelligence and facial recognition, CyberLink has developed the FaceMe® Facial Recognition Engine. Powered by deep learning algorithms, FaceMe® delivers the reliable, high-precision, and real-time facial recognition that is critical to AIoT applications such as smart retail, smart security, and surveillance, smart city, and smart home.

For more information on FaceMe®, please visit https://www.cyberlink.com/faceme/ 


Press Contacts

Corporate: Richard Carriere at press@CyberLink.com 
CyberLink Japan: Hiroyuki Imazawa at contact_pr_jpn@CyberLink.com or by phone at +81-3-5875-6650
CyberLink USA: Kelsey Duke at contact_pr@cyberlink.com or by phone at +1 (646) 571-0120
: Stefan Winter at contact_pr_deu@CyberLink.com  or by phone at +49-531-3873316
France: Laëtitia D'Urso at contact_pr_fr@CyberLink.com or by phone at +33-(0)1 79 30 90 39
United Kingdom: Richard Carriere at contact_pr_uk@CyberLink.com