Research On Iris Recognition Algorithm

With its uniqueness,stability,easy acquisition and high security,human iris is considered as the most research and market potential biometric technology.It has been widely used in national defense,financial industry and entrance guard system.This paper mainly improves the algorithm in three parts: iris location,feature recognition and matching,and achieves better experimental results.The main work and results of this paper are as follows:1、In view of the low efficiency and accuracy of the traditional iris location algorithm,a simulated annealing genetic algorithm is proposed for iris localization.In order to avoid the low efficiency of the genetic algorithm and easy to fall into the local optimal solution,this paper introduces the simulated annealing algorithm into the genetic algorithm,and combines the advantages of the strong local search ability of the simulated annealing algorithm,and sets the adaptive cross probability and mutation probability,and improves the probability of convergence to the global optimal.When the iris image is processed,the area growth method is used to remove the light spot.Then the inner and outer edges of the iris image are roughly detected by the Canny operator.16 pixels are extracted from the inner and outer edges of the rough location.The simulated annealing genetic algorithm is used to optimize the target function,and a set of optimal solutions is obtained.2、In feature extraction,this paper proposes an algorithm combining differential box counting with missing items.The dimension of the difference meter box can reflect the roughness of the texture in the iris image,and reflect the possession of the texture space.The gap can be used to quantify the cracks or gaps on the surface of the image.It is independent of the fractal dimension,and the combination of the two can better describe the iris texture information.When the feature is extracted,the iris image is normalized and enhanced,and the feature matrix of the difference counting box dimension of the iris texture and the missing item is extracted,and the feature extraction is realized by comparing with the threshold.3、In matching recognition,this paper combines the advantages of Support Vector Machine(SVM)and Hamming Distance(HD),and uses the feature matching recognition algorithm based on the combination of SVM and HD to match the pattern.The experiment shows that the algorithm has improved the location speed,accuracy and feature extraction,and has achieved good results.Compared with the traditional matching recognition algorithm,the algorithm has a higher correct recognition rate and is in line with the requirements of most occasions in practical applications.