Palm vein recognition system using multiresolution transforms

Abstract

Authentication is the process of verifying the identity of the user. newlineEarlier, for a person s identity verification, methods such as something to newlineremember, like a password, and something to carry, like ID cards were used. newlineThese older authentication methods have problems like remembering and newlinehacking of the passwords, and the ID cards can be easily stolen, lost and newlinemanipulated so that the security is compromised. To overcome these issues in newlineearlier authentication methods, the next higher level of security strategy, newlinebiometrics was introduced. Biometrics is the way of automatically identifying or newlineauthenticating the person based on what he/she possesses. It involves newlinephysiological and behavioural traits such as fingerprint, face, hand geometry, newlineiris, retina, vein (finger, palm and hand), speech, gait, key stroke, hand writing, newlineand so on. Each biometric trait has its own advantages and disadvantages. newlineThe palm vein trait includes more advantages like hygiene, user acceptability, newlinehighly secured being internal to the body, and liveness detection. Hence, the newlineproposed work uses palm vein trait for human authentication and identification. newlinePalm vein authentication can be employed in applications such as finance, newlineattendance for a person, access control, health care and library access. newlineUnderstanding the benefits and applications of the palm vein trait, newlinethis research proposes optimum solutions for automatic identification and newlineauthentication of human beings by incorporating different multiresolution newlinetransforms such as curvelet transform, and adaptive Gabor filter which describe newlinethe images at different resolutions, do not miss any of the discriminating features newlineat any scale and orientation, and provide faster computation. The performance of newlinethe proposed works is measured in terms of recognition rate, Equal Error Rate newline(EER), and identification time. The work is carried out with CASIA newlinemultispectral palm print database V1.0 and VERA palm vein database. newline