Real-time Iris Segmentation based on Image Morphology

Abstract

This paper introduces an efficient iris segmentation approach for unconstrained images. Proposed technique is robust to occlusion, specular reflection, variation in illumination and non-centered gaze. For pupil localisation, the input iris image is binarised using an adaptive threshold determined based on number of connected components. Further,pupil center and radius are obtained using spectrum image based approach. The proposed technique performs accurately (>97%) with low computation (<0.4 seconds/image). It has been observed that the proposed approach can be deployed to real-time biometric systems where time as well as accuracy cannot be compromised.