Iris recognition based on dynamic radius matching of iris image

This paper presents iris recognition method based on dynamic radius matching of iris image. First, the iris images are segmented to remove the eyelashes and eyelids. Then the individual feature of the iris image can be extracted by expanding their polar images into Fourier series. The obtained Fourier coefficient is used as the individual features for iris recognition. Moreover, in order to reduce the fluctuation caused by size of pupil and iris, the dynamic radius matching is introduced to calculate the similarity between the iris images. Experimental results were performed on CASIA V1.0 [6] public iris database having 756 iris images from 108 persons. The obtained accuracy rate was 94.8%.