An Improved Iris Segmentation using Pupil Candidate Bank for Non- Cooperative Iris Recognition

Abstract

This study focuses on iris recognition; the most accurate method of biometric identification comes from the person's unique characteristics and the permanence of the iris texture. A modified method in which a more precise segmentation procedure is added to an already existing efficient algorithm, boosting iris recognition's overall reliability and accuracy, has been tested with two datasets MMU and CASIA. However, MATLAB software has been used in this segmentation. Besides, this study focuses on the pursuit of iris segmentation in non-cooperative and the effect of noise situations because of the occlusion that challenges iris segmentation. Detecting eyelashes and shadows in the iris border may not correspond to a circular shape. Recent improvements have several enhancement methods for analyses, including the method Pupil Candidate Bank (PCB). While these approaches increase the accuracy of the original iris recognition system, they revise segmentation techniques in the integro-differential operator and Hough Transform. (PCB) the method could reduce the input quality; inaccurate iris segmentation may result in poor recognition. The enhanced iris segmentation approach has increased the overall accuracy of 98.9% of iris identification by improved performance to validate the authenticity of a subject.