“The company says that its technology’s advancement is the product of machine learning techniques, and that it will extend benefits to all of its SDKs and its entire Ver-ID line of applications.”
Applied Recognition has reached a new benchmark in its facial recognition technology. The Toronto-based company says it has attained a 0.2 percent Cross-Over Error Rate (CER) in National Institute of Standards and Technology testing with the FERET dataset.
The CER refers to equanimity between an identification technology’s False Acceptance Rate and its False Rejection Rate, which is “the best summary measure of face recognition technologies’ efficacy,” according to a statement from Applied Recognition. The company says that its technology’s advancement is the product of machine learning techniques, and that it will extend benefits to all of its SDKs and its entire Ver-ID line of applications.
It could help the company to find new partners and clients as well, and Applied Recognition appears to be targeting the financial services industry in particular. In a statement announcing the new benchmark, Applied Recognition Co-CEO Ray Ganong said the technology’s “new level of accuracy supports a completely automated workflow for financial transactions, such as opening a bank account or applying for a loan.”
August 31, 2017 – by Alex Perala