Iris Recognition Algorithm Analysis And Research

Biometrics is attracting more and more attentions and market demands,and has been researched and applied.Iris recognition technology is one of the technologies in the human biometric recognition technology in favor.Because of its unique superior physiological structure,such as:uniqueness,stability,security,making it an important means of identification of indentities in today's society,and it also has a wide range of implement prospects in defense,security,e-commerce,finance and other fields.In this paper,the iris recognition system is deeply researched and analyzed.A four-direction scanning method is proposed to locate the inner and outer edge of the iris quickly and accurately.Bipolar coordinate normalized iris area and the division of the module.The improved 25-Dimensional Gabor filter feature extraction;improved Ferns classifier method for iris feature training and sample matching test.This paper presents an improved algorithm mainly in the iris image preprocessing,feature extraction and pattern matching module of the three parts of the study,the specific details of the narrative:First,pretreatment:iris location is the key part of the iris image preprocessing module.In this paper,the method of fast and precise location of iris inner and outer edge based on four-direction scanning method is adopted.Firstly,the edges are extracted based on the improved Canny operator and the wavelet transform to obtain a closed inner edge.Then,the pupil is segmented by four-way scanning.After the edge points are re-determined,the radius of the inner edge is calculated.The edge of the inner edge of the center radius of the radius of the outer edge of the outer edge of the radius,and finally according to differential,integral outer edge of the precise information to obtain accurate positioning iris inner and outer edge of the iris texture feature for the more pure information,feature extraction,the need to do The preprocessing includes image normalization and image enhancement.In this paper,normalized bipolar coordinate normalization is used,and then the module is divided to select more and more pure iris texture information as possible.Equalization enhancement processing.Second,the feature extraction:feature extraction module is based on 5*5 direction to improve the use of two-dimensional Gabor filter.The feature of image is extracted from the selected iris region,then the features of the Gabor filter are reanalyzed and parameterized.The feature extraction and fusion of 25 texture directions are carried out,and the feature points are calculated and the feature codes are obtained.After treatment,the relevant effect diagram.Third,the pattern matching:based on Ferns(fern)classifier method.This paper introduces the classification of training set and test sample,the principle of classifier,analyzes the advantages of iris matching and compares it with support vector machine classifier,and compares it with the commonly used Hamming distance method.The ROC performance curves of the method are compared.In order to verify the algorithm proposed in this paper,we use CASIA and British bath two iris database to test the performance of more than 40,000 samples.The operation platform of this algorithm is implemented on the MATLAB software platform.