Cognitec upgrades FaceVACS biometric engine for higher identification accuracy with masks
Cognitec has upgraded its FaceVACS facial recognition engine with advanced deep learning methods and a new face finder, giving the new version higher biometric accuracy rates for masked faces, “strong poses” and exaggerated mouth expressions, according to a company announcement.
The biometric engine now has a detection rate for faces mostly covered with masks ten times higher than before, due to extensive research and development efforts on the FaceVACS face finger, Cognitec says.
The company offers a specialized version of its algorithm which consumes less memory for Linux-on-Arm and Android platforms, and this algorithm has particularly high matching accuracy for images of people wearing masks in Cognitec’s testing.
“Cognitec’s algorithm development is responding to our customers’ critical need for accurate identification and verification products in the current pandemic environment,” states Cognitec Managing Director Alfredo Herrera.
The company is now preparing its next biometric matching algorithm release, which is expected to have higher accuracy for general matching tasks, and for partially occluded facial images.
Worldreach recently chose Cognitec facial recognition to enable remote customer onboarding.
Several other biometric algorithms have been updated to better match masked faces, with SAFR’s the most recent example.