Image Histogram: Preliminary Findings of Anti-Spoofing Mechanism for Hand Biometrics Acquisition

Abstract

Biometrics data are prone to spoofing activities especially on its sensor levels where fake biometrics data can be generated to imitate genuine biometrics data. Fake biometrics are false biometrics data that resemble genuine biometrics characteristics. If fake biometrics is accepted by a biometrics system, the possibility of personal information and data to be stolen is high. The consequences lie in the unwanted access, and the public may become insecure to use biometrics as an authentication tool. Biometrics acquisition process with an added detection mechanism can help distinguish between genuine and fake biometrics data. It is possible by the use of near-infrared (NIR) light during acquisition process because the interaction between NIR light with human skin and fake biometrics are different; due to the living trait property possessed by a human. This paper shares preliminary findings of image histogram for both genuine and fake biometrics images acquired by NIR illumination. Observation on the image histogram reveals that there are differences to the image properties that can be used to distinguish the genuine and fake biometrics data. The approach can be extended to its usage as a detection mechanism for other biometrics data as well. The main principle lies in the difference of image response between genuine and fake biometrics data acquired by the NIR illumination.