Local pattern-based descriptors for iris recognition: A comparative analysis

Human iris has found widespread utility in person authentication problems. Reason behind this is the availability of unique texture patterns present in the human eye. Given the success of local pattern-based descriptors in texture classification, this paper aims to bring together different descriptors, based on local neighborhood of a pixel in an image. A comprehensive comparative analysis of 11 different descriptors is provided in this paper. All descriptors are evaluated under the same experimental setup using the IITD iris database. A critical analysis is presented under both the verification and identification modes of the biometric recognition system.