Study On Evaluation Method Of Visible Iris Image Quality On Mobile Terminal

With the rapid development of mobile smart terminals and the Internet,more and more functions are being implemented on mobile smart terminals.The use of biometrics to implement identity authentication for mobile smart terminals is one of the current study hotspots.Among them,iris recognition is an important choice for mobile terminal identity authentication because of its good activity and anisotropy.In traditional iris recognition,the iris images were taken under infrared light.However,infrared light not only causes eye damage,but installing an infrared light camera in a mobile smart terminal also increases production costs.Mobile iris recognition uses the camera embedded in the mobile smart terminal to photograph the iris under the visible light.It leads to the position and angle of the photographer and the camera are not fixed,the lighting conditions during the shooting are not controllable,the photographer's head or hand is shaken,etc.Thus,the iris texture in the iris image captured under visible light is occluded or unclear,and an iris texture that use for iris recognition cannot be extracted.Therefore,the evaluation of visible light iris image quality is introduced,and the image with severe interference of iris texture is excluded,and the kind and degree of iris image interference are analyzed to determine whether the image can be used for identification.In the existing visible light iris image quality evaluation,the Hough transform is used to locate the iris first.However,the shooting environment in the mobile iris recognition system is not controllable,and it is impossible to precisely locate the iris in the iris image whose quality is seriously disturbed.Therefore,this thesis designs a coarse-to-fine quality assessment method.While evaluating the iris image quality,the iris is gradually and accurately positioned,and the iris texture quality is classified and evaluated.The main contents are as follows:(1)Discuss the research status of iris recognition and quality evaluation at home and abroad,and use subjective evaluation method to analyze and classify the iris images in the library,and build a classification library for the algorithm test of this thesis;(2)In the coarse quality evaluation stage,Viola-Jones algorithm and FT method were used to perform a coarse selection on the visible iris image,and the region of the eye that did not conform to the Haar-like feature or the domainant region was not in the eye position due to severe interference makes iris images excluded.Then,the eye areas were located in nonseverely interfering iris images.(3)In the fine quality evaluation stage,a method of locating the iris by using the color and shape features of the iris is proposed to continue locating the iris region in the eye region.Then,design two critical quality assessment factors in the visible light iris image to quantify the quality of the iris area.Using the above two quality assessment factors as criteria,the fixed threshold method and support vector machine(SVM)method were used to classify the iris images,and the quantitative evaluation results of iris texture quality were obtained.The performance tests of the above methods are compared with the classification results of the subjective evaluation method.The results show that the proposed method can correctly evaluate the quality of iris images,eliminate the images that cannot be used for iris recognition,reduce unnecessary computational burden on the mobile terminal,and effectively improve the efficiency of iris recognition on the mobile terminal.