Skip to main content
— Presentation at ICCE2017 —
February 17, 2017
ICCE2017 (International Conference on Consumer Electronics) is the IEEE Consumer Electronics Society's flagship conference and held in Las Vegas on January 8 to 10, 2017. ICCE is co-located with CES (Consumer Electronics Show) that is global consumer electronics and consumer technology tradeshow. There were 139 oral presentations, 58 poster presentations and 5 keynote speeches. I made a presentation titled "Walkthrough-Style Multi-Finger Vein Authentication" that realizes accurate verification of a walking person (Photo 1).
Photo 1 Presentation scene
Recently, large scale acts of terrorism have occurred all over the world, thus strict personal identification is needed at sport stadiums and any other facility where large crowds gather. Strict identification takes a long time for large groups trying to access a venue. To solve the problem, real-time identification with high accuracy is needed. Biometrics is a technology for uniquely identifying humans on the basis of intrinsic physical features and has no risk of being forgotten or lost and is receiving a lot of attention.
Fig. 1 Configuration and appearance of
Fig. 2 Overview of authentication system
We proposed walkthrough-style multi-finger vein authentication method that achieves high accuracy and high operability of being identified while walking. Finger vein authentication technology has an advantage of being highly resistant to forgery or spoofing because finger veins are inside the body and invisible from the outside. Proposed user-friendly user interface enables users to place fingers while walking easily (Figure 1). To improve authentication accuracy when deterioration of accuracy occurs due to variation in position or posture of fingers, we use multi-finger vein patterns obtained at the same time (Figure 2). In finger vein capturing, each light source is controlled in accordance with the position of fingers detected by the depth camera for capturing multi-finger vein image clearly. Moreover, to achieve highly accurate finger vein authentication, a precise finger-vein region detection combining near-infrared (NIR) and depth cameras are proposed. We developed a prototype and authentication system for practical use and demonstrated that the proposed method provides real-time authentication with high accuracy and operability.
Future work will involve an approach for commercial applications such as authentication with moving fingers and fingers in various postures, and for the evaluation of large-scale data.
(By MATSUDA Yusuke)