Effective approach for iris localization in nonideal imaging conditions

Abstract

This paper presents “effective approach for iris localization in non-ideal imaging conditions”. We proposed an effective algorithm based on the color clustering method for coarse iris localization of non-circular iris boundaries and contrast enhancement used for pupil extraction. The specular reflection removal is carried out by morphological and gradient based method. Also circle correction and non-circular boundary detection is proposed. Upper and lower eyelids boundaries are detected and eyelashes are also removed based on average intensity analysis. The proposed method has been evaluated on the UBIRIS 2.0, NICE.I and NICE.II iris databases. It detects exact iris boundary and removes all irrelevant parts such as eyelids, eyelashes, reflection and gives more accurate results than other existing methods.