Novel iris location algorithm based on the structure of human eyes

Iris location is an important part of iris recognition, and speed of location and robustness of the location algorithm are factors which we should pay especial attention to. A novel iris location algorithm was proposed based on the analysis of the structure characteristics of human eyes. A point in the pupil was gotten by gray summing operator. Three inner boundary points which were not on the same line were found by the boundary detection template to locate the iris inner circle rapidly, and four outer boundary points that were not on the same line were found by the boundary detection template to locate the outer iris circle. 756 iris images comprising an iris image database were located. Our algorithm only costs 1.59min, average time of location reaches 0.126sec and accuracy of location reaches 99.07%. This algorithm is much faster than any other classical methods such as Hough transform and the circle detection operator at speed. Furthermore, it ensures the accuracy of location and the robustness.