Fujitsu unveils integrated palm vein and facial biometric system for wallet-free payments
Fujitsu has announced the development of an integrated non-contact palm vein and facial biometric authentication technology for retail payments.
In use cases such as ATMs, in which palm vein authentication is used to identify a person from among several tens of thousands of enrolled users, a second factor, such as a card, is used to narrow the field for efficient matching, Fujitsu says. Adding facial data that can be captured with a typical camera enables the same efficiency without requiring a card or other input, even for use cases with one million-user scale volume.
Fujitsu says it has developed an algorithm to reduce the processing load of extracting facial angles and expressions to one-tenth of the size, to enable instant matching results. The user’s facial data is captured while the payment terminal is operated, and narrows the field of possible matches. The palm scan performed at the end of the process authenticates the user, and is supplemented by facial image matching if necessary, for instance if a portion of palm vein data is not captured. The load on the authentication server is limited by separating the processing of the two modalities for high-speed matching without a major increase of computing resources.
The technology can also be used for other high-volume use cases, such as event admission, according to the announcement.
Details about the technology will be shared at the International Conference on Image Processing (ICIP) 2018 in Athens, Greece on October 7, and Fujitsu plans to make it available in fiscal 2020.
Palm vein authentication technology from Fujitsu was recently adopted by the Japanese Racing Association to authenticate users at self-service betting kiosks. Trials of facial recognition payment systems have also been recently launched in settings with limited user volumes in Korea and Finland.