New active contours approach and phase wavelet maxima to improve iris recognition system

Due to the randomness of iris patterns, iris recognition systems are the most accurate, reliable and efficient way to recognize and identify people. However, the complex structure of an iris image results in the difficulty of iris representation especially for iris images of insufficient quality. In this paper we propose a new active contours models applied to segment iris images based on active contours without edges to localize the pupil and GVF snake combined with Wavelet maxima multiscale edge detection to localize the outer iris boundary, after normalization step using Daugman method (homogeneous rubber sheet model) a new technique for feature extraction based on phase wavelet maxima is used to extract iris features, evaluation tests with Casia v3 database confirmed the efficiency of the proposed methods for iris segmentation and feature extraction, The experimental results have shown that the proposed system could be used for personal identification in an efficient and effective manner (accuracy over 99%) and comparable to the best existing iris recognition systems.