Research On The Algorithm Of Iris Recognition System

The personal identification has been widely used in various fields in modern society. Traditional means of recognition is unable to meet the needs of modern society.The emerging biometric recognition develops quickly with their own unique advantages.Compared with other kinds of biological recognition technology,Iris Recognition System is quite suitable for identification because of its uniqueness,stability,accuracy,rapidity,collection and non-invasion and so on.An ideal Iris Location Algorithm should have both a relatively high recognition rate and processing speed.However,the current researches on Algorithm of Iris Recognition are not enough to meet the two requirements at the same time.Therefore,it is necessary to conduct a research to explore it further.The key problems of improving iris recognition technology are how to locate iris correctly, and extract texture features of iris image effectively,as well as design efficient classifiers.The present study mainly explores the improvement of recognition rate and processing speed of Iris Localization Algorithm.Based on the analysis of traditional means of recognition,it makes researches on iris localization,features extraction and template matching.At first,due to the fact that Iris Location Algorithm is time-consuming,with low accuracy rate,the study proposes a novel algorithm for iris localization which can remove interference effectively and improve iris location's speed rapidly.Then this kind of algorithm divides iris location into two parts,which are inner boundary location and outer boundary localization.As for inner boundary localization,based on the binaryzation processing of pupil area,the author makes use of the method of morphological opening operation for elimination of noise in order to locate the center and the radius of the pupil through gray projection.As for outer boundary localization,the author adopt morphological closing operation to remove ample texture in iris area and design a model of edge detection, which searches for 4 boundary spots in external boundary to locate the center and the radius of the iris.Besides,against the existing deficiency of extracting method,the present study put forward a method of using local gray minimum points to find out the feature points in effective iris region, depending on the length and direction of texture to get iris binary code by abandoning the fake feature points.Then according to the code,the study makes statistics on a large number of training samples with Hamming-distance classifier to get the distribution of Hamming distance respectively from the same and different iris for exploring classified criterion of pattern matching. At last,the feasibility and accuracy of improved algorithm has been verified through experiment results.All the algorithms has been tested in CASIA standard iris database.The experimental results indicate that the iris recognition system has been improved in the study,the processing speed and recognition rate have increased significantly,and the system has a very good effect on recognition.