Implementation of Wavelet Transform-Based Algorithm for Iris Recognition System

The purpose of this study is to design a system that will capture the iris image and develop a reliable feature extraction algorithm for iris recognition system. The proposed system is a complete iris recognition system with hardware and software components in which the focus is on the implementation of algorithm based on wavelet transforms. The system consists of the video camera that is interfaced through a frame grabber using the MATLAB program to capture an image of the human eye. The camera includes adjustable chin support, NIR filter and NIR diodes for the lighting and distance settings. The algorithm implemented in software performs segmentation, normalization, feature encoding, and matching. The feature encoding is performed by decomposing the normalized 24 x 240 pixels iris image using Haar and Biorthogonal wavelet families at various levels. The vertical coefficients are encoded into iris templates and stored at the database. The system is evaluated in two modes: verification and identification. The HD values are used as threshold levels to identify the iris image. The number of degrees of freedom is calculated for inter-class comparisons. The test results at different coefficients show that in terms of efficiency, the Haar wavelet decomposition at level 4 is the highest with a Correct Recognition Rate (CRR) of 98% at a feature vector length of 120 bits. The Equal Error Rate (ERR) of the system is 2%. The metrics show that the proposed system provides highly accurate recognition rates and suggest the most appropriate choices that need to be made for best results.