An online biometric authentication system based on eigenfingers and finger-geometry

An approach to personal authentication using the fusion of the finger-geometry and the novel biometric features called eigenfingers at the score-matching level is presented in this paper. The online biometric system integrates finger-geometry features extracted from the four fingers and eigenfingers features extracted by means of the Karhunen-Loève (K-L) transform applied to the four finger strip-like regions. Additionally, the system has a liveness detection module, which uses an IR image of the dorsal surface of a hand. Authentication experiments were conducted on a database consisting of 1270 hand-images (127 persons). The verification results, EER = 0.04% and minimum TER = 0.04%, suggest that the system can be used in medium/high-security environments.