An efficient interior and exterior boundary detection-based two level iris segmentation

Abstract

Iris recognition system's performance relies upon the method of segmentation of iris from the eye image. The segmentation process of iris still stumbles upon with few tricky challenges, particularly in separating the iris from eye image. The prevailing eyelids and eyelashes in the image leads to the lessening of the accuracy. A two-level segmentation' methodology was proposed. Initially, the iris image was converted to greyscale image and normalised. The adaptive median filter (AMF) was utilised to remove the noise. The noise removable image was segmented using two-level segmentation method. In interior boundary segmentation section, the image was segmented utilising some methods like Gaussian pyramid, anisotropic diffusion, thresholding, centroid computing, polar transform, and radius computing. Exterior boundary segmentation section performs zigzag collarette process. Finally, the IBS was subtracted from EBS; to give the segmented iris. Experimental results when compared with ACWOE and K-means demonstrated the superiority of the proposed technique.