An Improved Multi Biometric System for Authentication

Abstract

Biometrics is the new technology for calculating and measuring the body parts of a person. It is playing an important role in identifying an individual. It signifies a metrics related to person characteristics (physiological or behavioral). Biometric system may be based on single modal or multiple modals. Multimodal system also termed as multi-biometric system (hybrids two or more modals) are becoming popular. The idea behind the paper is to implement and improve the authentication processusing multiple trait for identifying a person. Here, the combination of Iris & Fingerprint based biometric model is presented. The relevant features (key-points) are extracted from these two traits in parallel and then they are passed through matching module. The key-points are extracted using Discrete Wavelet Transform (DWT) and Speeded-Up Robust Feature (SURF) descriptor and then passed to the matching (mapping) module where mapping is done using the Normalized Weighted Sum-Rule. The experimental result showed that the proposed multi-biometric model performs well showing clear variation in FAR and FRR against the existing models.