A comparison of fused segmentation algorithms for Iris verification

Abstract

Recent studies show fusion at level of segmentation to be useful for more robust iris recognition rates compared with simple segmentation. In this paper we perform Sum-Rule Interpolation at level of the result of the normalized segmented iris images using the well-known Daugman’s algorithm, since the process of normalization is essentially composed by two parts: Iris segmentation, in which the pupillary and limbic polar curves are detected and Iris normalization: a normalized representation of the iris texture is created using angular and pupil-tolimbic radial coordinates. For evaluation we propose an experimental fusion scheme using three automatic segmentation algorithms which have reported good results and are not computationally expensive. The experiments were performed on the CASIA V3-Interval, CASIA.V4-Thousand and UBIRIS V1 datasets showing increased recognition accuracy for representative feature extraction algorithms.