A Novel Registration Approach Based on Curve Fitting Technique for Iris Images

Iris registration aims at processing many iris images which collected from the same person, these images are different in pupil size, texture and so on. The study in this research field is very valuable. In this paper we have proposed that apply curve fitting technology in iris registration. First, the Harris corner detection has been employed to check the feature of iris images, followed we have combined the local jet and LBP approach to obtain some potential matching points between iris images, then according to the ratio of the Euclidean-distance from the matched points to the center of each pupil we have drawn the plot. Finally we have done the curve fitting, using the curves to estimate transform model and acquired the iris registration images. The experimental results have shown that our method is more accurate and more close to source image than traditional method which found in current literature.