Performance Analysis of Iris Recognition System Using DWT, CT and HOG
S J Ajmi, Prof. Arun P S · IOSR Journal of Electronics and Communication Engineering · 2016
Authentication and recognition can be achieved using biometric methods such as fingerprint, DNA, Iris etc. Human iris is the most reliable biometric because of its uniqueness, stability and noninvasive nature.The iris recognition system includes image acquisition, pre-processing (normalization and segmentation), feature extraction and recognition.For feature extraction, different methods considered are DWT, CT and HoG.2D DWT, used to create a feature vector is decompose up to its 8 th level and the approximate coefficients are retained to represent the feature vector.Another transform technique namely contourlet transform (CT), which is a new 2 D extension of wavelet transform, employs the combination of laplacian pyramid and directional filter bank.The feature vector build using CT reflex the directional information of iris.This iris recognition also perform feature extraction using HoG.Whose feature vector is computed by dividing an iris image in to blocks and there after computing the gradient followed by histogram.The database used is CUHK and CASIA.These methods are evaluated based upon False Rejection Rate (FRR) and False Acceptance Rate (FAR).