The Quality Evaluation Research Of Iris Image

In the recent years, a new biometrics identification—iris identification have received more and more attentions for its stability, reliability and no violation. During the image acquisition process of an automatic iris recognition system, unqualified iris images are probably rejected, and raise the error reject ratio. In this dissertation, much research works have been done in the iris image assessment, and four performance, Image Definition, valid iris ratio, elliptic ratio and same centre ratio, also have been sent out.The dissertation introduces a new method of Image Definition. Using Fourier transform to find image's frequency in frequency fields. The more frequency is the higher frequency fields, the higher Image Definition is. So the dissertation gives a new performance a to measure how much frequency in the higher frequency fields.The dissertation introduces a new method to find how much the iris in the image. By analyzing the distribution of the Histogram, one can easy find the difference of the iris and other interferences.The dissertation introduces a new method to check whether the iris is regulation. If an iris image is not ordinary, its elliptic ratio and same centre ratio which calculus from the centre of inner and outer iris's circle and radius.The dissertation introduces a new method to search the inner side of the iris. It combines the Histogram analysis and Hough transform, and improve the iris system's reliability.