Chinese University researchers achieve 99.15% facial recognition accuracy
A team of researchers from the Chinese University of Hong Kong has reportedly achieved an accuracy rate of 99.15% with its facial recognition algorithm.
According to a report in ChinaInternetWatch, these results are based on a test using the Labeled Faces in the Wild database. This database has been specifically created for studying the problem of unconstrained facial recognition, and is seen as a benchmark for these systems.
Called DeepID, this new algorithm is reportedly more accurate than Facebook’s DeepFace, based on its LFW score. The team of researchers are also responsible for another two facial recognition algorithms that the ChinaInternetWatch report says are also more accurate than DeepFace.
Reported previously in BiometricUpdate.com, Facebook launched DeepFace earlier this year.
Xiaou Tang, the professor who leads the team of researchers at the Chinese University of Hong Kong plans to make facial recognition technology free for Android, iOS and Windows Phone developers through the production of a new inclusive SDK.
According to recent research, the global facial recognition market is being driven by demand from government and finance.