An Efficient Iris Recognition System Using Integer Wavelet Transform

Abstract

An iris recognition mechanism is being discussed in presented paper, which is based on Integer Wavelet Transform (IWT). UBIRIS. v2 iris database is works as the input image for implementation of this mechanism. For efficient iris region localization combination of Total Variation Model and Hough Transform is used. An iris recognition algorithm is developed by using four level integer wavelet transform (IWT). IWT is an interesting alternative of discrete wavelet transform, which gives less computational complexity because it gives integer value as the coefficient. Four level IWT is firstly applying into the segmented iris image, which generates 256 sub-bands. But only 192 sub-images are taken into consideration after ignoring the high frequency sub-bands. After that the energy of 192 sub-bands is calculated for feature extraction, which helps to generate a 192 bit binary code. For recognition of individuals, normalized hamming distance is computed between the trained and input binary codes of 1-D feature vector. Matching results have been plotted and discussed in the paper.