Detection of Non-iris Region in the Iris Recognition

In view of iris disturbance of eyelid, eyelash and reflection spot affect the accuracy of iris recognition, a detection approach of non-iris region in the iris segment is proposed. Based on mathematical morphology knowledge, the proposed method uses the least squares combined with edge detection to determine eyelid occlusion, then, eyelash and reflection spot can be detected with threshold method. The experimental results show that this method may detect disturbance accurately and effectively, and can exclude it from the iris region, so that the effect to recognition can be suppressed in the iris segment. At the same time, it reduces the complexity of the algorithm greatly, which indicates that it is an effective and feasible method of non-iris region detection.