Iris localization algorithm based on gray distribution features

Influenced by the noises, for example, eyelid and eyelash occlusions, iris localization algorithms are difficult to keep a right balance between real-time quality and accuracy. Therefore, a novel iris localization method with effective noise eliminating is proposed in this paper. For the iris´ inner edge location, morphological open is used to eliminate noise based on separating pupil region by binaryzation. Then the pupil´s center and radius are located accurately by gray projection. For the iris´ outer edge location, morphological close is proposed to eliminate the rich texture within the iris area. Based on this method, an edge detection template to search four directional points within a small scope of the probable boundary is designed and thus the iris´s centre and radius can be located. Simulation results show that, compared with the pre-algorithms, the new algorithm not only greatly improves the accuracy and the speed, but also shows good robustness.