FacePhi has modified its facial recognition algorithm to help identify people wearing masks. The new software is 99 percent accurate, and specifically focuses on the area around the eyes to minimize the threat of failures and false positives.
The technology was developed in response to the ongoing COVID-19 pandemic, which has encouraged people to wear masks to promote safety and public health. FacePhi’s full-face algorithm is still the default for users of its facial recognition platform, but people who are wearing masks can voluntarily choose to activate the new system to improve its performance.
FacePhi believes that the new system will be particularly beneficial to people in the healthcare industry, where hospitals and clinics will be able to continue using facial recognition for access control and patient identification. The technology allows healthcare facilities to replace physical health cards and sign-in sheets with more sanitary contactless alternatives.
“The use of this technology goes beyond COVID-19,” wrote FacePhi in a statement. The company recommended its technology for any “environments where the use of a face mask is constant and contactless identification solutions [are needed] in order to prevent contagions.”
FacePhi has already deployed a contactless authentication system at the Kangbuk Samsung Hospital in South Korea through its partnership with NSSMART.
“This is preventing fraud in health insurance and offering a quick experience that greatly limits contact, something that will mark the future of healthcare,” continued FacePhi.
FacePhi is best known for its SelphID solution, which has gained considerable traction with financial institutions throughout Latin America. Some of the company’s newest clients include Banco de Corrientes and Santander Argentina, in addition to Interbank in Peru. All three organizations are using SelphID to facilitate the mobile onboarding of new customers, allowing people to sign up for new accounts with a selfie and a photo of their official government ID.
–
(Originally posted on FindBiometrics)
Follow Us