Template security in iris recognition systems: Research challenges and opportunities

Abstract

An iris recognition framework is a standout, stable, secured, and widely accepted biometric in the secured authentication infrastructure. It has been broadly perceived as one of the grounded biometrics credited to its high precision execution. In this paper, a comprehensive analysis of several iris template security techniques has been carried out together with open opportunities for further research challenges. The analysis mainly pointed out that majority of the techniques results in a trade-off among various performance and security parameters, e.g., FAR, FRR, EER, revocability, diversity, and security. The well-known open iris databases such as CASIA, IITD were accessible to complete the experiments. It also has been noticed that the template security scheme for uni-biometric system may not work for multi-biometric system. Among the two broadly available template protection techniques, feature transformation schemes outperform its counterpart cryptosystems mainly due its unlinkability and irreversibility properties.