NEC face recognition engine conducts rapid crowd video capture
NEC’s NeoFace Accelerator is a PCI Express low profile board with rapid face authentication capabilities based on 4K video acquisition and face matching with an existing photo database, according to a report by eeNews Europe.
Drawing up to 25W and measuring 68.9×167.65mm, the board can track many individuals from a crowd of hundreds as they walk throughout large venues, transit stations or shopping centres.
NEC says its face recognition engine is powered by Intel Arria 10 field programmable gate arrays (FPGAs) operating on Intel Xeon processor–based servers to increase the software’s performance to a level where a person can be identified seamlessly from a high-resolution photo containing dozens of faces.
“Facial recognition in a moving crowd requires highly advanced techniques when compared to still images because these cameras are affected by many factors: camera location, image quality and lighting, along with the subject’s size, walking speed and face direction,” said Tadashige Kadoi, general manager of the IoT platform development division at NEC Corporation. “Intel FPGAs and their parallel processing capability help NEC to enable fast and accurate collection and processing of images from even 4K high-resolution remote cameras.”
NEC NeoFace outperformed its competitors, specifically in face-in-video evaluation, in nearly all the tests conducted by the U.S. National Institute of Standards and Technology (NIST) in March.
The tests assessed the accuracy of the face recognition technology in two real-life test applications including entry-exit management at an airport passenger gate, in which it determined how well the engine could recognize people as they passed through an area one at a time without stopping or looking directly at the camera.
NEC’s face recognition technology ranked first, achieving a matching accuracy of 99.2 percent and an error rate of 0.8 percent, which was less than 25 percent of the second place error rate, according to Intel.
The second test assessed the technology’s ability to detect suspicious individuals at an indoor stadium, in which the test subject was standing far away from the camera.
In addition, the person was instructed to frequently change the direction in which they were facing.
Once again, NEC’s technology came in first place with an error rate half that of the second place error rate.
For the NeoFace Accelerator vehicle, NeoFace facial recognition engine software IP is integrated into an Intel Arria 10 FPGA.
Intel teams also collaborated with NEC to improve the performance of NeoFace data center server technology.
NEC NeoFace Accelerator includes the Intel Arria 10 FPGA, as well as an Intel MAX 10 low-cost FPGA and Intel Enpirion power devices.
Previously reported, a pilot project started in early June on flights from Dulles International Airport to Dubai is testing NEC’s commercial-off-the-shelf NeoFace Express device and cloud-based NeoFace matching as key components of a frictionless biometric exit process.