Real ID advanceswith Optical Character Recognition (OCR) and cross-racial facial recognitionaccuracy

SINGAPORE - MediaOutReach - 10 December 2018 - The recent spate ofdata breaches--including a flawin Google+ that led to its shutdown and the reported hacking of 120 million Facebook users' data--have highlighted the abuseof digital identity.

While having a digital identity brings convenience, thechallenge is to ensure that individuals have full control and protection oftheir data. Technologies for safeguarding digital identities have quicklyemerged in the U.S. and Europe, but developments have been slow in Asia due to itsdiverse language mix and cross-racial facial features, says Real ID's CEO ChongHan Ping.

Real ID, whose partners include Alibaba-backed SenseTime, recently launched its screening and face-learningtechnology. Its AI and blockchain technology shortens the onboarding and KYC(Know-Your-Client) process for remittance transactions by as much as 90% and significantly reduces human error, says Chong.

"Multi-language verification is rare, and we havefocused our efforts on building partnerships in different markets to improveour language-learning and OCR technology as well as to shorten machine-readingtime," he adds. OCR is the recognition of printed or written text characters by a computer.

Real ID, whose employeesinclude former data scientists and machine-learningexperts from Tencent, China National Research Centre, and Nanyang TechnologicalUniversity, is in talks with authorities in India, Thailand,Philippines, and Indonesia to build a regional digital ID ecosystem for frictionlesscross-border business.

Real ID adopts SenseTime's facial recognition technology,which is empowered by livenessdetection, face matching and machine learning to eliminate any deception. The use of blockchain technology ensures that recordscannot be altered retroactively without the consensus of the network. A combinationof both on-chain and off-chain transactions also allows verification results tobe audited by different parties, without the need to reveal underlying identityinformation. For furthersecurity, the data is backed by validators who are authorised to issue a truststatement on behalf of the user.