Abstract:

The authors proposed a combination of Fourier descriptors with the least square method for iris localization. On locating the inner boundary of iris, an improved tracking algorithm was introduced to obtain pointsequences by tracking the edge of the binarization pupil. After Fourier descriptors transforming, the least square method was finally used to fit the inner circle. On locating the outer boundary, the gray gradient information on the left and right parts of the iris was appropriately used. Based on this information, the outer boundary pointsequences were extracted. Similarly, the outer circle can be fitted by the least square method at last. Experimental results show that the method can overcome the eyelashes and eyelid occlusion, and also improve the loc
alization speed and accuracy. The results can meet the requirements of realtime iris recognition system.

Key words: iris localization, Fourier descriptors, least squares method, contour tracking