Efficient Iris Recognition Based on Optimal Subfeature Selection and Weighted Subregion Fusion

Algorithm 1

Function CFS.

Input: 𝑀: number of keypoint, 𝐷: dimensional number of feature element.

Output: Discriminative sub-feature 𝐶^𝑗_𝑖.

(1) Begin

(2)    Setting the number of finite-delete-sorting based on orientation at 𝑛;

(3)    Setting the number of finite-delete-sorting based on magnitude at 𝑚;

(4)    For 𝑖=0 to 𝑛

(5)     Getting the sub-feature 𝑂^𝑠_20−𝑖 from all keypoints’ feature element;

(6)     For 𝑗=0 to 𝑚

(7)      Getting the sub-feature 𝑉^𝑠_16−𝑗 from sub-keypoints corresponding 𝑂^𝑠_20−𝑖

(8)      Calculate classification accuracy rate of 𝐶^16−𝑗_20−𝑖 based on SVM classifier;

(9)      Save the related parameters, which are 𝑖, 𝑗, and CAR𝐶^16−𝑗_20−𝑖;

(10)     End for;

(11)    End for;

(12)  Getting the MaxCAR𝐶^16−𝑗_20−𝑖 and its corresponding sub-feature element;

(13)  End.