Model for iris texture deformation with changing pupil size

Abstract

Iris images have been proposed as a means of biometric verification and identification. Transformation of iris image coordinates to account for change in pupil size were performed by Daugman (Daugman 1993) based on a linear model. In order to investigate the effect of change in pupil diameter on the iris texture, we have used an experimental set-up to capture iris images by step change in light intensity. Accordingly, a non linear model is proposed to account for the deformation of the iris texture. Genetic algorithm is then used to determine the optimal values of the deformation model parameters to minimize the difference between the captured iris image and the stored iris image. These parameters might need to be incorporated in the code of the iris in order to allow for variations in pupil size due to either light intensity change or accommodation.