Mitigation of Contact Lens, Eye Surgery, Pupil Dilation, and other Challenges on Iris Recognition

Iris biometrics offers the potential for highly accurate identity verification. However, over 25 million people in the United States and over 125 million people worldwide wear contact lenses, and it would be impractical to require users to remove their contact lenses for recognition purposes. As a result, developing methods to measure and mitigate any degradation associated with contact lenses and other atypical eye conditions has become increasingly important to the future of iris recognition technology. In this DHS SBIR research project, we will study, analyze and mitigate the potential adverse effects of contact lenses, eye surgery, and varying pupil dilation, template aging, and sensor interoperability, on iris recognition. We will collaborate with Clarkson University to leverage their existing data collection efforts to examine potential performance variations of iris images with and without different types of contact lenses, before and after surgery, or at varying pupil dilation. After data collection, we will conduct a detailed analysis of the specific problems. If, as expected, significant performance differences are observed, image processing algorithms and machine learning approaches will be developed to detect and mitigate instances in which matching performance appears to have been degraded.