The development of face recognition payment system in CAS

Abstract: Online payment no longer enter the password, verification code, direct face card. Chinese Academy of Sciences, Chongqing Green Intelligent Technology Research Institute in the first development of the face recognition payment system, so that online payment only use mobile devices to take a picture. Face Recognition payment system is

Online payment do not need to enter the password, verification code, directly brush the "face card." Chinese Academy of Sciences, Chongqing Green Intelligent Technology Research Institute in the first development of the face recognition payment system, so that online payment only use mobile devices to take a picture. Face recognition payment system is a payment platform based on face recognition system, which has been introduced in foreign countries in 2013. Last year, the Chinese Academy of Sciences, Chongqing Green Intelligent Technology Research Institute Intelligent Multimedia Technology Center launched the payment method of research.

Online payment do not need to enter the password, verification code, directly brush the "face card." Chinese Academy of Sciences, Chongqing Green Intelligent Technology Research Institute in the first development of the face recognition payment system, so that online payment only use mobile devices to take a picture.

Face recognition accuracy rate is up to 99.8%.

Face recognition payment system is a payment platform based on face recognition system, which has been introduced in foreign countries in 2013. Last year, the Chinese Academy of Sciences, Chongqing Green Intelligent Technology Research Institute Intelligent Multimedia Technology Center launched the payment method of research.

According to Zhou Xi, director of the center, the Center and the United States University of Illinois image generation and processing laboratory, the National University of Singapore to build, in the face recognition system has a relatively mature technology. He told reporters that their global first face data acquisition array, 91-point face synchronous collection, has formed 50 million face of the world's largest human face training database in Asia. At the same time, a visual object analytic framework based on depth learning is established, which effectively improves the stability and accuracy of face recognition in dynamic environment, and can optimize the recognition effect of the changeable illumination, multi-angle, occlusion and so on.

"Face recognition is not only accurate but also fast." "Zhou Xi said that their face recognition system in the United States Carnegie Mellon International Standard face Recognition test library to achieve a 99.8% recognition rate, while the international other algorithms the best result is 97.6%." At present, their face recognition system has been used in border control station automatic clearance system, dynamic face recognition attendance machine, multi-attribute dynamic face recognition system.

On this basis, they have completed the research and development of the key technology of face recognition through 1 years.

It only takes 1 seconds for the system to recognize the payer's identity.

Zhou Xi said that face recognition is a biometric identification technology based on human facial feature information, and the biggest feature of technology is that it can avoid personal information leakage and identify it by non-contact way. Because the human face is not replicable, compared to fingerprint identification is not easy to replicate, there is no password to forget, leakage and other risks, so security is higher.

"It only takes 1 seconds for the system to recognize the payer's identity," he said. "The center is currently developing a face recognition mobile payment app that will require users to install the app on a terminal (cell phone, tablet) and link a bank account or credit card to a face recognition payment system," Zhou Xi. When used, users only need to take a picture of the camera, the system in 1 seconds to complete identity verification, to achieve fast payment, the entire payment process is expected to take only 4-6 seconds.

The system analyses the face of the database and collects 2 million facial features for combinatorial analysis, and has formed an intelligent analysis algorithm. Ordinary people can not distinguish between the twins, the system is able to distinguish one eye, so the possibility of brushing the wrong face is very small. At the same time, the system can be detected in vivo, using photos or video to "Brush", at a glance will be identified by the system.

In addition, the system has the function of online learning, each time brushing the face, will automatically learn the body's variable characteristics, such as long fat or thinner, users can not worry about their fat and thin will affect the system recognition.

Mobile payment app launches next year

"The recognition mobile payment app for face recognition mobile payment system is expected to be launched in the second half of 2015." "Zhou Xi Introduction, at present they are and part of online payment companies, electric dealers to discuss cooperation, when users pay online shopping only download terminals, according to the program to shoot all angles of the photo settings, you can enjoy a convenient and quick" face Card "payment method.

Zhou Xi revealed that next, the center will continue to develop face recognition mobile payment terminals, by the time the merchant uses a face recognition mobile payment terminal with a camera at the cashier's desk, the user only needs to take a picture of the camera, and then the system searches the database for the face and matches it with the stored payment information. And finally accurate into the user account, the merchant and the user to confirm the amount of consumption after the realization of rapid payment. In the real sense to achieve without card payment, so that customers do not have a wallet shopping, and may even replace the existing POS machine.