Research On Algorithms Of Iris Localization

Iris recognition authentication technology is a high-end research in the field of secure identity authentication, bioinformatics identity authentication technology is a never backward and eternal high-end issue in this field. Iris had attracted more and more researchers’ attention rely on its own unique physiological characteristics and unique geometric properties. Iris authentication have non-intrusive, stability and security so that it set off another round of the boom in the biological information recognition research field. And this wave has set off the academic enthusiasm of the domestic and international researchers. With the development of society as well as in-depth research work, the characteristics information of iris of the authentication has been widely used in coal, banks and other daily life.Iris image preprocessing is a powerful basis for identity authentication, iris localization is an important part of the iris preprocessing, so the pros and cons of the iris localization is the assurance to achieve efficient and reliable identity authentication. Iris location is the center of pre-processing stage, and the targeted research is the focus of this article. Meanwhile, timeliness and rate of iris localization will directly affect the entire identification process, so it is the top priority how to improve the efficiency of time and locating that decided the efficiency of recognition.Next, the work of this article will be briefly introduced:1. In order to achieve in-depth research and analysis of research more classic localization algorithms in the field, such as: these pre-processing algorithms proposed by Dr. Daugman, Dr. Wilde, Dr. Tan and Dr. Feng were depth theoretical analyzed and implemented, summed up the advantages and disadvantages that each method exists, and presented current issues which prevalent and difficulties unresolved in the research. The analysis found that too large search area is the common problem. Some improvement were made on the solutions of this problem in this article, the advantages of Wilde’s and Daugman’s were combined, the improved algorithm based on small search area was proposed.2. Introduced the new method. This algorithm searched the main parameters of inner and outer circle of the iris according the main features of two circles. The process of the algorithm is that: searching the point of the pupil, opening and closing operation de-noising, and the inner and outer circle’s localization. In the first, finding a random point in the pupil. Then the go around accorded known point to locate the coarse boundary of the circle. In the second, using the matching circular template and Hough change to achieve internal and external circular locating. The algorithm does not search the approximate location of the pupil in original image, instead of detecting it according the known point of the pupil. This method greatly reduced the targeted search area, which affect the calculated amount of search and shorten the positioning period.3. Completed the experimental verification of the performance of the algorithms. In the designed experiments of the article, firstly selected the two iris databases(CASIA-Version 1.0 and CASIA-Version 3.0_interval). Then randomly and respectively selected 50 class samples, and each class included at least seven images. Set the desired algorithm experimental environment, described the implementation principle of every step. According to the setting algorithm, quality assessment, localization, normalization, enhancement and characteristics processing method to operate the database. During the experiment, not only use the same type of hardware equipment, but also set the same software algorithm environment to ensure the rigor of the comparison among different algorithms.In this paper, differently compared and analyzed performance from two aspects, the localization time and the recognition rate. The experimental results confirmed that this described locating method not only has a significant increase in speed, but also has improved in recognition accuracy.