NEC and The Face Recognition Company each upgrade biometric systems accuracy for masked faces
NEC Corporation has launched a new biometric facial recognition engine for improved accuracy when matching people wearing masks.
The company’s facial recognition technology was already relatively tolerant of masks, according to the announcement. As facial recognition technology must adapt to the “New Normal,” NEC developed the new algorithm which specifically focuses on the (periocular) area around the individual’s eyes for use in a broad range of applications from smart phones to workplaces.
The engine performs mask detection, and based on the result selects which algorithm to use for biometric face matching. The engine was found more than 99.9 percent accurate in internal 1:1 verification tests with masks of various colors and patterns.
The new engine will be compatible with the company’s Enhanced Video Analytics product from November, and multi-modal Bio-IDiom Services.
The engine is being tested at NEC headquarters along with a thermal camera system.
FRC anticipated occlusion problem
The Face Recognition Company (FRC) has announced a system upgrade to enable the accurate biometric identification of people wearing masks, which the company says has a negligible performance difference for masked and unmasked users.
FRC had already anticipated the challenge of identifying individuals whose faces are obscured with biometrics before the COVID-19 pandemic began, according to the announcement. The mask wearing that has become normalized around the world due to the pandemic will likely remain for some time, FRC argues, based on historical evidence like the reinstatement of mask-wearing laws after their initial lifting during the 1918 flu pandemic.
The company says it is also committed to staying ahead of possible future threats as facial coverings become more sophisticated and incorporate elements designed to defeat biometric identification systems.
FRC came out of stealth late last year.