Panasonic’s new facial recognition engine scores most accurate in NIST testing
The core engine of Panasonic Business’ new deep learning facial recognition platform Face Server, developed jointly with the National University of Singapore, has achieved the most accurate results in industry testing by the National Institute of Standards and Technology (NIST), IFSEC Global reports.
The platform, which has already been tested and deployed at Tokyo’s Haneda airport, can identify faces at up to 45-degree angles to the left or right, and up to 30 degrees up or down, as well as faces partially hidden by sunglasses or face masks.
“Face Server changes the game in terms of detection accuracy,” said Gerard Figols, Category Manager at Panasonic Security. “Not only does it maintain 90%+ accuracy for faces that are partially covered by sunglasses or face masks, it also recognises faces from photographs that are up to 10 years old.”
Face Server can process data from up to 20 cameras per server in real time, and search up to 30,000 registered reference faces at high speed. It can be used with Panasonic’s range of iPro Extreme cameras to automatically adjust settings for optimal image capture with “iA (intelligent Auto) mode” and “best shot feature” capabilities. The iA function also reduces server and network load by analysing images on the camera and transmitting only the best quality ones to the server.
It was recently announced that face recognition server software from Panasonic will be used for biometric access control at the stadium for Belgian football club R.W.D. Molenbeek.