 |
| Masks aren't an obstacle to NEC's facial recognition system |
The Japanese company NEC has introduced a face recognition system that can recognize people even when wearing a mask and adapt it to the new standard, as face masks have become an essential form of preventing the spread of the Coronavirus.
The technology company is developing a system to meet the needs of allergic patients who wear masks when the coronavirus epidemic causes development to speed up.
“With the emergency caused by the Coronavirus going on for a long time, demand has continued to grow and we have already adopted this technology ... market,” Shinya Takashima, assistant director of digital platforms division at NEC, told Reuters.
The system detects that the person is wearing a mask and polishes exposed parts such as eyes and around them to verify the identity of the person, and the user has previously recorded pictures of his face.
According to NEC, verification takes less than a second and claims the accuracy rate is over 99.9% and the system can be used for security doors in office buildings and other facilities.
NEC declined to disclose the price of the system, and its biometric and video analytics business (including facial technology systems) aims to generate 100 billion yen ($ 970 million) in sales during fiscal year 2021.
Takashima said: The new system was launched in October and was sold between the customer (Lufthansa) and Swiss International Airlines.
Takashima explained that facial recognition technology means no need to present a security card that can be lost or stolen, and that technology can also help people avoid touching surfaces in various situations.
He said: Because of the impact of the Coronavirus, contactless testing has become critical. We hope that we will be able to contribute to achieving security and protection by intensifying efforts in this field.
UK municipal police are using NEC's NeoFace real-time face recognition technology to match the faces of a crowd of faces on their watch list.
According to a report by the National Institute of Standards and Technology, the facial recognition algorithm failed to identify 20% to 50% of photos of people wearing masks before the Coronavirus pandemic, but the accuracy had improved significantly by the end of 2020.