The Research Of Iris Image Rotation Problem

With the development of the internet and communications technology, information security have become increasingly demanded. Because traditional identification technology has been unable to meet the requirements of society, the new biometrics algorithm emerged. Iris recognition is in this context to develop. As the iris physically advantages,such as unique, stability, non-invasive, make iris recognition become a research hot spot.Iris recognition system includes iris image acquisition module, pre-processing module (the iris location, the normalized iris image, iris image enhancement), feature extraction module, pattern matching module. The iris image will directly affect the recognition results. In the shooting environment, due to external factors such as the impact of the iris to obtain the deflection angle of the image that may occur, which will directly affect the recognition results. Classic recognition algorithms in dealing with a larger deflection angle of the iris image, have some flaws. This major work is the above solution to the problem. Consider of a rotation invariant in Zernike moments, so we use Zernike moments to amend the iris image.The Zernike moment of order n, with repetition m for a continuous image function f(x,y) that vanishes outside the unit disk iswhere n is a non-negative integer and m positive and negative integers subject to the constraints n-|m|even and|m|< n,p is the length of vector from the origin andθthe angle between vector and x axis in counter-clockwise direction.{Vnm(x,y)} is a set of basis plural function which domain in x2+y2=1,and they are orthogonal. Rnm,are the Zernike radial polynomials in polar coordinates defined asSuppose the image after a rotation of a is fα(ρ,θ), then the new Zernike moment fα(ρ,θ) is,Ifαis the angle of rotation,Znmα, is the Zernike moment of the rotated image and Znm is the Zernike moment of the original image, the absolute of two Zernike moments are equal. This is the rotation invariant of Zernike moment. [Zernike moment Znmis a plural, then ifm=0, Znm=Znmα, if m≠0,If the original iris image is f(x,y),and the image after rotation is fα.Then,we get m values of angles.These m angles is 2π/m less than the nest ones. In this paper, we use the spindle of the iris to determine a unique rotation.The classic spindle direction is:We get m spindle direction and control the amend spindle direction and the spindle direction of the matched image in a same interval(such as (-π/m,+π/m)), then we get the unique angle of rotation.As it can be drawn, when the image occurs a rotation, if the original image Zernike mo-ments Z(nm) is known,we only need to calculate the new Zernike moments after the rotation Znm',and then make a division for the precise degree rotation and finally amend to the image by. using the polar coordinates transformation methods.In this paper,we have the experiment in the CASIA 1.0 based on given methods.We find that using Zernike moments do overcome the impact cased by the deflection of the iris image. The rate of recognition in this paper is a little higher than classical method.Because the calculation of Zernike moments is complicated, system uptime used in computing the Zernike moment take up a lot of time.Therefore, the calculation of Zernike moments will be used largely determined the efficiency of the whole system. The method of calculation used in the experiments is relatively simple and fast method of computing Zernike moments are emerging, which will need our further study.