A longitudinal study of iris recognition in children

Biometrie recognition is dependent on the permanence of the biometrie characteristics over long periods of time. However, there has been limited research in this area, particularly in children during development. This paper presents the start of a longitudinal study of irises in children to understand when biometrics can be used reliably and the effect aging has on the biometric modality as a child grows. Data was collected and analyzed in children ages 4-12 years over three visits, spaced approximately six months apart. This is one of the few iris collections spanning this broad age range in children. The results show that there is a slight decrease in match scores between the resultant comparison of collection 1 to collection 3(12 months difference) and the resultant comparison of collection 1 to collection 2 (6 months difference); analysis shows this difference is not statistically significant. Additionally, the data analyzed resulted in very similar iris recognition performance when examining a subset of subjects in fifth grade and a subset of subjects in first grade. These results could indicate that the iris biometric characteristic is stable over time, at least as early as age 4, the youngest group tested in this work. Additional longitudinal data is needed to support this hy-pothesis.