A Fast Iris Location Algorithm Based on Minimal Variance Searching Center
Xiao Kun-ping · Computer Knowledge and Technology · 2010
In order to improve the real-time characters of iris identification, this paper presents a new fast iris location algorithm. Firstly, the iris image was denoised and the threshold segmentation was carried out by applying the OSTU method. Then the rough center point and radius of the inner edge of iris were obtained by using the projection method. At last, according to the obtained rough center point and radius, the inner edge of iris was located accurately by applying the presented algorithm in this paper. For the outer edge’s location, based on the prior knowledge and the gained circle parameters of inner edge, the extra edge points and noise points were removed and the searching range was reduced, and then the accurate center point and radius of the outer edge of iris was gained by using the same way with the inner location. Experimental results show that this algorithm can determine the inner and outer edge of iris quickly and accurately, the location speed is nearly ten times faster than those traditional algorithms, and reduce the searching blindness.