Enhancing Security of Multimodal Biometric Authentication System by Implementing Watermarking Utilizing DWT and DCT

Abstract

Conventional personal identification techniques for instance passwords, tokens, ID card and PIN codes are prone to theft or forgery and thus biometrics isa solution thereto. Biometrics is the way of recognizing and scrutinizing the physical traits of a person. Automated biometrics verification caters as a conducive and legitimate method, but there must be an assurance to its cogency. Furthermore, in most of the cases unimodal biometric recognition is not able to meet the performance requirements of the applications. According to recent trends, recognition based on multimodal biometrics is emerging at a greater pace. Multimodal biometrics unifies two or more biometric traits and thus the issues that emerge in unimodal recognition can be mitigated in multimodal biometric systems. But with the rapid ontogenesis of information technology, even the biometric data is not secure. Digital watermarking is one such technique that is implemented to secure the biometric data from inadvertent or premeditated attacks.This paper propounds an approach that is projected in both the directions of improving the performance of biometric identification system by going multimodal and, increasing the security through watermarking. The biometric traits are initially transformed using Discrete Wavelet and Discrete Cosine Transformation and then watermarked using Singular Value Decomposition. Scheme depiction and presented outcomes justifies the effectiveness of the scheme.