Camvi announces record biometric accuracy in UMass Faces in the Wild benchmark
Camvi Technologies has announced it achieved record-breaking accuracy of 99.87 percent in a recent University of Massachusetts’ Labeled Faces in the Wild (LFW) benchmark test.
The result indicates the effectiveness of Camvi’s algorithm for unconstrained images, as does the most recent Face Recognition Vendor Test (FRVT) report from NIST in the Wild Photos category. The company says unconstrained images are the most challenging for face biometrics, as they come from real-world sources such as surveillance and photojournalists’ footage.
LFW has been one of the best-known and most respected benchmarks in the field of face recognition, according to the announcement, since it was introduced in 2007 to stimulate research in face recognition for images captured in common settings.
“Setting the world record in the LFW in addition to the FRVT results solidifies our leadership position not only in the U.S., but also in the world,” said John Chen, President and Chief Executive Officer of Camvi Technologies Inc. “While we are proud to have the world’s #1 ranking in the LFW as well as the FRVT results, we are also very pleased to deliver the record-breaking accuracy to our customers in our products at unrivaled speed and scalability. We relentlessly push the envelope in our quest for the most accurate, fastest and most scalable face recognition technology. We will continue to let our results emphasize our commitment.”
Camvi’s core technology is embedded in its CamviServer, CamviEdge, CamviClient and CamviAPI. The company showcased its technology at the 51st ASEAN Foreign Ministers’ Meeting in Singapore in August, and a company representative delivered a presentation on biometrics at the edge at FedID in October.