Biometric Aging Effects of Aging on Iris RecognitionSeptember 2014
Topics: Biometrics, Biosensors, Image Processing, Pattern Recognition
The fundamental question being asked is, does iris recognition performance remains constant and persist well over time? Or conversely, is the iris structure susceptible to irreversible changes that cause greater dissimilarity as time intervals between recognition events increase? Medical literature reports that the pupil size decreases slightly as people age . However, the iris is naturally protected within the body, with protective coverings (eyelids and cornea) that prevent the physical surface "wear and tear", and loss of moisture experienced by fingerprints and the skin and soft tissue on the face. There are certainly disease and vision related ailments that occur with age; the eye can suffer presbyopia, cataracts and other diseases. But the fundamental assumption for iris recognition is that the structure and patterns in the elastic connective tissue of the iris remain stable and permanent throughout life.
The belief that iris features are permanent and stable was asserted by the originator of the first patented iris recognition algorithm, John Daugman. As commercial iris sensors have only existed since the 1990s and significant collections for military or border management have only been around since the early 2000, there is limited longitudinal data has been available for study. Daugman's assertion has never been empirically validated against a large population observed over 5 years or more. This work is a step toward that goal.