An Efficient Iris Recognition Using Correlation Method

In this paper we propose the algorithm, which is different from Daugamn’s iris recognition algorithm. In the proposed algorithm the segmentation and matching method is quite different from Daugman’s method. We segment the image by considering only boundary of pupil. The localization of pupil can be done using thresholding and plotting virtual circle of proper radius can segment the region of interest closer to pupil boundary. Normalization is done using Daughman’s rubber sheet model. To remove the intensity variation due to varying lighting condition the histogram equalization is used. Matching has done by the correlation method. Since the two dimensional correlation gives the extent of similar distribution of pixel values, it can be used to identify that iris belongs to same person or not. Recognition algorithm is implemented in MATLAB software. An experimental result shows false acceptance rate (FAR) and false rejection rate (FRR) of system. According to the result we conclude that the proposed algorithm accuracy is good and it is efficient and adaptive to environment, it works well even for noisy environment, variable illumination.