Fast Iris Localization Algorithms

Iris recognition is a cutting-edge interdisciplinary research in the applied mathematics and the pattern recognition, Iris recognition is the frontier of biological characteristics identification field, as it has higher security and stability. Iris localization is that iris is segmented in the colleted image and iris recognition requires accurate iris localization for successful processing.At present, some scholars have proposed a number of more effective iris localization algorithms. However, the speed of iris localization algorithm is still the main difficulty in the iris recognition system. Furthermore, at the accuracy rate of localization, it is still the difficulty that boundary points extracted contained noisy points.The author presents a fast iris localization algorithm in this paper. Firstly, determine the effective area of the iris image by spot features so as to reduce unnecessary searches in non-effective region. In addition, when the iris boundary located by calculus detection operator needs a lot of computation, the author presents four-quadrants algorithm to improve calculus detection operator.At the same time accelerating the speed of iris location, it is necessary to ensure the accuracy of iris location. The author uses the clustering algorithm to remove noisy points.The results of the experiments carried out on database CASIA-IrisV3-Lamp and CASIA-IrisV3-Thousand with Matlab indicate that the proposed algorithm in this paper has a higher accuracy and robustness.