High-performance human identification method using iris image processing

With the increasing security needs, biometry is used as a solution to the recognition of people. This concept is based on structural and behavioral traits. The structural traits include fingerprint, face, iris, etc. The features of the patterns extracted from these traits are unique and relatively more stable and time invariant compared to behavioral traits such as signature, and handwriting, which change with time and are imitate. Characteristics related to the iris pattern have revealed very good capabilities compared to other types of these features. This paper presents and evaluates novel human identification method, which uses iris image processing. The proposed method consists of several subsystems, which include segmentation, normalization, feature extraction and the use of the codes comprising the features extracted from the iris for identification. This paper utilizes MATLAB and the image base of Institute of Automation, Chinese Academy of Sciences (CASIA), which includes 756 gray images of 108 people. The obtained results using suggested algorithms, ultimately led to a success rate of about 95.91% and average processing time of 0.73 s on the image base of CASIA. This indicates the high accuracy and speed of the proposed algorithm.