A Two-Way Image Quality Enhancement for Iris Recognition System Using Modified Enhanced Histogram Equalization for Normalization

The term “iris recognition” refers to a method of biometric identification and authentication that uses unique pattern recognition techniques for identifying individuals. Low-quality iris images pose a drawback in iris recognition system as the recognition of each individual depends solely on a good quality iris image. This paper aims to address the low-quality iris images with the following characteristics: poor resolution, low contrast and poor illumination at the normalization stage. It is proposed in this paper, a method that enhances the low-quality iris image at normalization stage. The proposed technique was developed using Matlab R2016a and it was tested using two standard datasets such as Chinese academy of science institute of automation (CASIA) and unconstrained biometric iris (UBIRIS). The finding of the paper reveals that the proposed technique has an accuracy of 96.3% as compared to existing techniques. It was also found that the average improvement for the false acceptance rate (FAR) is by 0.02%, while for the false rejection rate (FRR) improvement is about 0.05% which can be seen as an enhanced method compared to existing techniques. Based on the findings, it is clearly evident that the proposed approach possesses the capacity and the ability to enhance low-quality iris image at the normalization stage.