We present results of the first systematic study to investigate the degree to which template aging occurs for iris biometrics. Our experiments use an image data set with approximately four years of elapsed time between the earliest and most recent images of an iris (23 subjects, 46 irises, 6,797 images). We compare the match and non-match distributions for short-time-lapse image pairs, acquired with no more than 120 days of time lapse between them, to the distributions for long-time-lapse image pairs, acquired with at least 1,200 days of time lapse between them. We find no substantial difference in the non-match, or impostor , distribution between the short-time-lapse and the long-time-lapse data. We do fine a noticeable difference in the match, or authentic , distributions. We find that, for a fixed value of decision threshold, the false reject rate increases by about 50% for the long-time-lapse data relative to the short-time-lapse data. The magnitude of the increase in the false reject rate varies with changes in the decision threshold, and across the three algorithms that we studied. However, all three algorithms have an increased false reject rate over the range of feasible values for the decision threshold. Thus, our experimental results demonstrate that there is a template aging effect for iris biometrics.
Citation: IEEE Transactions on Pattern Analysis and Machine Intelligence
Pub Type: Journals