VinAI biometric recognition technology for phones works with protective face masks
Claiming it is one of the first to achieve successful results, the Vietnamese lab said it is making the system available for free, as part of its contribution to COVID-19 relief efforts.
To ensure stable accuracy, the technology leverages a deep learning model created by researchers from the Institute for research and application of optimization (VinOptima). The model is designed to recreate neural network activity in the human brain, and it automatically trains and collects facial information to identify people even when wearing masks, the company says.
VinAi claims it uses a different algorithm and tech deployment to provide better accuracy than the technology already available on the market. In this case, the recognition system uses normal camera sensors instead of infrared and depth sensors, which will keep the price low and make it easy to integrate with camera systems to track and authenticate employees.
“Masks are essential item in the COVID-19 prevention rules, but when users wear masks, the accuracy of current facial recognition technology can be reduced by more than 50 percent. Therefore, VinAI has urgently developed this technology to promptly meet social needs,” said Dr. Bui Hai Hung, Director of VinAI Research, in a prepared statement.
Additionally, VinAi has developed a tool to monitor people who are wearing masks to ensure social distancing measures are respected. The lab is already working with VinSmart electronic equipment manufacturing company to integrate the technology with Vsmart phones so users can use biometric facial recognition to gain access into their phones without removing the masks.
A number of biometrics companies have recently announced upgrades or new versions of their biometric algorithms to identify people wearing masks including SAFR, Innovatrics, Speech Technology Center, NEC, Telpo, Alibaba, ZKTeco USA, SenseTime, Herta, Corsight and Hanvon.