Dilation aware multi-image enrollment for iris biometrics

Current iris biometric systems enroll a person based on the best eye image taken at the time of acquisition. However, recent research has shown that simply taking the best eye image and ignoring pupil dilation leads to degradations in system performance. In particular, the probability of a false non-match increases when there is a considerable variation in pupil size between the enrolled eye image and the probe eye image. Therefore, methods of enrollment that take into account pupil dilation are needed to ensure reliability of an iris biometric system. Our research examines a strategy to improve system performance by implementing a dilation-aware enrollment phase that chooses eye images based on their respective empirical dilation ratio distribution. We compare our strategy of enrollment to that of the randomly chosen eye images, which is the current enrollment procedure for most iris biometric systems. Our results show that there is a noticeable improvement over the random scenario when pupil dilation is accounted for during the enrollment phase.