Iris recognition using 2-D Gabor filter and XOR-SUM code

Abstract

Iris based person identification has become very largely used biometric technology. This paper proposes a technique of feature extraction of iris images, which uses 2-D Gabor filter with four orientations. XOR and SUM operations on the real and imaginary output for each orientation result into a number which is further encoded into bits to give the feature vector. Experiments show that XOR-SUM Code is providing increased recognition accuracy (i.e. GAR) as compared to another feature extraction technique DCT. GAR value for XOR-SUM Code is 96.33% while for DCT, its value is 90.79%.