A methodological framework for investigating age factors on the performance of biometric systems

Any individual's biometric data are likely to change with the passage of time and, as a result, developing biometric applications for long-term use is a challenging task. One of the factors which increases the challenge of dealing with ageing effects in biometric systems is that the age of an individual is a continuous variable, and it is impossible to investigate and understand ageing issues in biometric systems other than using discrete age "bands". However, this division of a given population into age-bands has generally been a rather arbitrary exercise, making it difficult to optimise or compare results in different studies. In this paper, we will investigate, document and analyse the effects of age-band assignment, improving our understanding of how to manage age-related data and pointing to the possibility of more objectively determining optimal age-bands which offer a greater possibility of minimizing the sensitivity of a system which relies on such information. We study specifically the iris (physiological) and signature (behavioural) modalities.