NEC claims win in NIST iris biometrics accuracy test
New iris recognition accuracy benchmarking from NIST is out, and NEC has claimed the top accuracy rate, at 99.59 percent for 1:N identification for images with both eyes from 500,000 people.
The IREX 10 test conducted by the U.S. National Institute of Standards and Technology evaluated 9 iris recognition algorithms from seven developers, benchmarking biometric identification accuracy with both two eyes and one eye. Neurotechnology scored the most accurate result with one eye.
In recent years, NEC has used its artificial intelligence capabilities to refine its iris recognition technology’s performance with the lower-quality images commonly captured in operational environments, according to the announcement. This, the company says, has caused a significant improvement in the algorithms’ error rate.
NEC plans to develop ‘walkthrough’ style solutions with its iris biometrics, in combination with its face biometrics, for highly accurate multi-modal authentication. This approach will provide effective contactless authentication even when users are wearing masks and caps.
The company notes it has also scored high marks for accuracy in NIST face biometrics testing.
The importance of privacy and human rights considerations to the NEC Group AI and Human Rights Principles is also emphasized in the announcement.