Statistical Line Based Palmprint Recognition Using Elliptical Gabor Filters

Most of the previous research using palmprint as a biometric trait for personal authentication has concentrated on enhancing accuracy. In this paper, to speed up the recognition process, we propose a novel method using the statistical line based approach to improve the efficiency of the Palm Code. Normally, palm codes from different palm images are similar. The structural similarities between palmprints will reduce the performance of the palmprint identification system. Hence, to avoid the correlation between the palm codes, two elliptical Gabor filters with different orientations are used to extract the phase information, and two elliptical Gabor filters are used for the Fusion Code and the Orientation Code. After the Fusion Code and the Orientation Code have been obtained, they are fused to obtain a single feature vector, Palmprint Phase Orientation Code. The similarity between two palm images is measured, using the normalized hamming distance. Using the Hong Kong PolyU palmprint database our experimental results show that the proposed method gives a promising result.