Design And Implementation Of Iris Recognition System Based On Wavelet Transform And Log Gabor Filter

With the coming days of a world full of information,convenient and reliable Identity Authentication Technology(IAT)is becoming more and more important,however,the traditional IAT cannot satisfy the high demand for information safety and some new Biometric Identification Technology(BIT)emerges at this moment.Among all these existing BIT,iris recognition technology is regarded as the best and the most promising BIT with a better stability,accuracy and safety.Thus,iris recognition technology has become a hot research study in terms of recognition algorithms and application.Iris recognition system consists of four parts:iris image capturing,image Pre-processing(iris location and normalization),feature extraction and pattern matching.This paper studies the image preprocessing,feature extraction and pattern matching in iris recognition based on the analysis of the existing research results,and then utilizes the method of wavelet transformation and Log Gabor filter to extract iris features.Use the coarse to fine matching method to complete the recognition and the mainly research project is as follows:First of all,this paper locates the iris based on the different features on the inner and outer iris edge.To the inner iris edge,the image is firstly binarized using the threshold method based on the grayscale characteristics of iris image,then we can get the separated pupil after denoising based on mathematical morphology.Afterwards,the edge is extracted using Canny operator and the inner iris is located accurately through Hough transformation.On this basis,the outer iris edge can be located accurately through improved calculus which can help to avoid searching blindly and only to deal with the interested areas instead.The algorithm calculation in this paper is greatly decreased and the location speed and accuracy is greatly improved compared with the traditional location methods.Besides,in terms of iris features extraction and encoding,the iris recognition method based on wavelet transform was analyzed,and combining the features of iris texture,designed for Log Gabor filter in iris feature extraction,and then combine the advantages of the two algorithms,the iris feature extraction based on the method of wavelet transform and Log Gabor filter.The specific process is the first the iris image after pretreatment of the three layer decomposition using wavelet transform,global information extraction can effectively characterize the iris texture feature,and disperse these data to implement iris coarse classification,forming a small sample set,and then using the Log Gabor filter,local texture information of iris image is extracted after preprocessing,the formation of the iris feature template after the quantization encoding.Lastly,this paper utilizes a coarse to fine matching method in terms of iris matching and the rough classification of iris according to wavelet transform encoding results,based on the characteristics of Log Gabor encoding for fine matching,namely in the small sample set with the Hamming distance calculating the similarity of iris feature template to complete the identification.This method avoided the blindness of the iris matching process,improving the iris recognition accuracy and efficiency.The iris recognition algorithm in this paper uses the experimental samples from the CASIA iris database(V1.0),including 756 images from 108 subjects.The simulation experiment is performed on Matlab R2010b.The simulation result shows that iris recognition method based on wavelet transform and Log Gabor is effective in improving the recognition speed and accuracy.