Iris matching algorithm based on weighted Hamming distance
WU Youshou · Journal of Tsinghua University(Science and Technology) · 2009
An iris matching algorithm is developed based on the weighted Hamming distance using a multi-scale Gabor filter bank to extract the phase characteristics of the iris texture.The Gabor filters with different scales and orientations have different descriptive capacities,so the features they extract have different discrimination abilities.Unequal weights are assigned to the output of each filter according to these dissimilarities when calculating the Hamming distance,then the verification is made based on the weighted Hamming distance.The experimental results show that the equal error rate of the present iris recognition system is 0.47% comparing to the traditional algorithm based on the normalized Hamming distance,whose equal error rate is 0.97%,the recognition performance of proposed system is improved obviously.