Biometric Identification Systems: Feature Level Clustering of Large Biometric Data and DWT Based Hash Coded Bar Biometric System

Abstract

Biometric authentication systems are fast replacing conventional identification schemes such as passwords and PIN numbers. This paper introduces a novel matching scheme that uses a image hash scheme. It uses Discrete Wavelet Transformation (DWT) of biometric images and randomized processing strategies for hashing. In this scheme the input image is decomposed into approximation, vertical, horizontal and diagonal coefficients using the discrete wavelet transform. The algorithm converts images into binary strings and is robust against compression, distortion and other transformations. As a case study the system is tested on ear database and is outperforming with an accuracy of 96.37% with considerably low FAR of 0.17%. The performance shows that the system can be deployed for high level security applications.