An optimal strategy for dilation based iris image enrollment

Estefan Ortiz, Kevin W. Bowyer, Patrick J. Flynn

Abstract: The progression of research into understanding and mitigating the effects of pupil dilation on iris biometrics is at a point where a formalization of the problem is necessary to tie together several research directions and results. Past research has shown that differences in dilation in a (probe, gallery) pair lead to an increase in false non-match rates. Additionally, analysis continues to show that there is at least an approximate linear relationship between increase in dilation difference and degradation in match scores. Lastly, dilation-aware based enrollment techniques have shown to be a promising approach to addressing matching errors due to pupil dilation difference. This paper establishes a framework based on an assumed linear relationship between match scores and dilation difference and shows that the optimal image to enroll based on pupil dilation is the image which has a dilation value near the mean or median depending on the measure of dilation difference.