Core demands
When "QR code" payment is popularized to pancake stalls at street corners, "face brushing payment" and other "face walking" methods are becoming more and more common "street smart scene". Baidu Brain and its senior partner, Boya Yingjie, hope to jointly create a 3D "face brushing" traffic equipment for Jinan Metro, so as to enable the 3D face recognition algorithm in the urban rail transit industry for the first time, realize the full face brushing traffic, completely get rid of the embarrassing situation of forgetting to bring money, cards and mobile phones, and make entering the station faster.
Solution
Before passengers take the bus using face painting for the first time, they need to verify the identity of "face picture+name+ID card number" through mobile phone APP. At the same time, the public security verification API interface is called online to complete real name authentication with the online live detection API interface.
When passengers "brush their faces and enter the station", Boya Yingjie's intelligent face brushing traffic device can automatically project 18000 speckles through the built-in 3D camera module, which can quickly collect complete three-dimensional face data, and complete offline 3D live detection with Baidu Brain's face offline recognition SDK multi-modal live detection algorithm. At the same time, the detection data is transmitted to the local server, and a complete face recognition service has been deployed - "Baidu Brain Face Recognition Privatization Deployment Scheme", to complete the whole process of face recognition. The fully localized algorithm service effectively ensures the privacy and security of personal data.
In the documentary "Crossing the City" broadcast by CCTV, a pair of twin sisters verified the recognition of the device and sent a "challenge" to the subway "face recognition" - the sister completed the verification of "face online calling identity" with mobile app and successfully entered the station through "face brushing", but the sister who has no obvious difference in appearance with her sister was mercilessly "rejected" by the gate, The accuracy of its algorithm can be seen.
At present, each station of Jinan Line 1 can enter the station with a face, and one channel is set for each gate group. The upcoming Line 3 plans to increase the proportion of face recognition channels to 50%, and Line 2 may be applied to all gates.
