Comparative study of biohashing technique using different feature extraction methods

Abstract

A comparative study of biohashing technique using different feature extraction methods has been carried out. A basic biohashing technique involves two steps. First is the extraction of the features from the biometrics and second is the discretization of the obtained features with the random numbers. In this paper, two different data bases and two different feature extraction techniques have been studied. The first data base consists of self generated images and images from ORL and Indian face database and the second data base is taken from the standard Yale-B data base. The two different feature extraction techniques that have been used are the optics based joint transform correlator technique and the principle component analysis technique. Hamming distance has been used to discriminate the different populations. The performance of the techniques has been analyzed by using the false rejection ratio (FRR), false acceptance ratio (FAR) and the equal error rate (EER) in order to compare the effect of the different feature extraction methods on the biohashing technique.