The Research Of Fast Clarity Evaluation Algorithm In Iris Recognition System

In recent years, with the rapid development of computer and Internet, how to protect national security and personal information not to be violated became an issue. The security of information is closely related to everyone, such as bank account and various accounts on the Internet, which represent the personal identity, concern about the security and benefit of oneself. If the account information is stolen, the consequences could be a disaster. The traditional identity recognition methods, such as ID cards keys e-cards and passwords, are very likely to be forged, be stolen, and be decoded, which are unable to satisfy the requirement of information security. On this occasion, biometric recognition is showing the way.Biometric recognition takes the advantage of one’s physiological characteristics or behavior characteristics to identify different people. The uniqueness of physiological characteristics and behavior characteristics make the biometric recognition has a high degree of security and accuracy. Among all of the biometric recognition, Iris recognition is one of all the biometric recognition who has the best performance benefits the lowest error rate, which is widely used in the area of financial business national security and so on, according to its high level of uniqueness stability and accuracy.Iris recognition system consists of four part, they are iris image acquisition, iris image preprocessing, feature extraction and matching identification. Its specific steps are as follows:first, take iris images by specialized image capture device; second, preprocess iris image, which include image clarity evaluation, iris segmentation, and normalization; the purpose of this part is to remove the poor quality image, therefore improve the accuracy of the whole system; then extract and quantify the iris feature, save them in the form of number; the last part, matching the iris feature of template library with the testing iris. Therefore iris image clarity evaluation is one of the most important parts of the iris recognition system, which will exclude the iris image of poor quality, thus improve the efficiency of the whole system indirectly.Most of traditional iris image quality assessment algorithm is complex, which takes a long time for evaluate process. In order to improve the efficiency of the iris recognition system, combining with the practical requirement, this paper devotes itself to the research of proposing a new fast image clarity evaluation algorithm. The work of this paper focus on the following aspects:Firstly, we have carefully studied the common algorithms for image clarity evaluation, In this paper of the iris recognition system, Based on the results of previous studies on the iris image clarity evaluation algorithm research and improvement, a fast algorithm for iris image clarity evaluation. Through carefully studying the common algorithms for image clarity evaluation, this paper propose a fast iris image clarity evaluation based on selecting region of interest (ROI). There are two methods to select the region of interest, one is based on the empirical research choosing the center part of the whole image as the ROI, the other one is based on the rough location choosing the ROI. Then evaluate the clarity degree of the ROI of iris image.In this paper, the experiments are based on database of CASIA V2 and V3 iris image library. The experimental results show that the iris image clarity algorithm proposes in this paper can quickly and accurately differentiate the iris image of different clarity.