SAFR from RealNetworks has released an updated version of its flagship facial recognition system. SAFR 3.4 arrives a few months after the October debut of SAFR 3.0, which enhanced the platform’s ability to identify people who are wearing masks.
The latest update, on the other hand, will improve SAFR’s passive liveness detection capabilities. The system needs less than 0.3 seconds to confirm that a real person is present, with a 95.27 percent accuracy rate. It also lets administrators set their preferred risk threshold, and can be configured to send an alert whenever someone attempts to spoof the system.
The other major feature being introduced with SAFR 3.4 is an SMS Watchlist Alarm. The utility will send a notification whenever a specific set of conditions is met, as is the case when a particular individual arrives, or when someone tries to enter a restricted area. However, the SMS Watchlist Alarm is only available for Windows customers at the moment.
SAFR 3.4 can be integrated into an existing security setup, and generates an auditable record of authentication events. The improved liveness system will function for both masked and unmasked individuals, and does not require the user to nod, blink, or perform some other action to demonstrate their presence.
“The new version of SAFR adds significant new levels of security for access control, whether it is authentication for physical access, or validation of a person’s identity for eKYC purposes,” said SAFR Product Management Senior Director Eric Hess. “SAFR works with 2D RGB video streams and dramatically lowers the costs and complexities associated with hardware based solutions.”
RealNetworks recently released a free mask detection app that runs on SAFR’s computer vision tech. Setelsa Security and a federal agency in the United Arab Emirates have both integrated the company’s technology into their own access control systems in the past few months.
April 8, 2021 – by Eric Weiss