Iris localization using Hybrid Algorithm containing Circular Hough Transform, Fuzzy Clustering Method and Canny Edge Detector

Abstract

Iris segmentation comprises of a sequence of operations including initial Iris and pupil detection as a most prominent region of image with pronounced round shape, outlining outer iris border and final refinement of visible iris part by rejecting regions occluded by reflection spots, eyelids and eyelashes. The results are mask of iris region i.e. set of pixels, which are visible points of iris together with its inner and outer borders. This paper presents an efficient hybrid algorithm to segment iris in unconstrained environment where human recognition is developing for images which can be captured without asking humans i.e. CCTV surveillance etc. by removing noise such as eyelashes and eyelids. It is a challenging task to get proper iris region from input image to get it recognized from trained dataset of the individuals. The proposed algorithm gives high accuracy rates in classification and matching.

Keywords: Circular Hough Transform, Fuzzy Clustering, edge detector, sclera region