Skip to Main Content
Continuous efforts have been made in searching for robust and effective iris coding methods, since Daugman's pioneering work on iris recognition was published. However, due to lack of robustness, the error rates of iris recognition systems significantly increase when images contain large portions of noise (reflections and iris obstructions), resultant from less constrained imaging conditions. Current iris encoding and matching proposals do not take into account the specific lighting conditions of the imaging environment, decreasing their adaptability to such dynamics conditions. In this paper we propose a method that, through a learning stage, takes into account the typical noisy regions propitiated by the imaging environment to select the higher discriminating features. Our experiments were performed on two well known iris image databases (CASIA and UBIRIS) and show a significant decrease of the error rates in the recognition of iris images corrupted by noise.