Personal identification using local Gaussian quadrature filter pair phase quantisation of hand vein images

Abstract

Hand vein pattern is a biometric feature in which the actual pattern is the shape of vein network and its characteristics are the vein features. This paper proposes a new approach and uses local phase quantisation with Gaussian quadrature filter pair for hand dorsal vein identification. The proposed work extracts the phase information computed locally in a window for every pixel position by employing Gaussian quadrature filter pair. The phases of six frequency coefficients are quantised and it is used to form a descriptor code for the local region. Whitening transformation is used to decorrelate these local descriptors and a histogram is generated for every pixel which describes the local pattern. The proposed work is experimented with minimum distance classifiers and results are examined for recognition rate, false acceptance rate (FAR), false rejection rate (FRR) and equal error rate (EER). Experiment results show that the proposed work outperforms the existing methods.