The Research Of Fast Quality Evaluation Algorithm In Iris Recognition System

In recent years,with the rapid de,velopment of Internet and information technology.people's lifestyles have taken place great changes and filled with convenience and intelligence,and the hidden dangers of the infornation security are gradually increased.which have cauaht the attention of the society.How to protect people's information security has become the research hotspots.Traditional identification methods,sruch as ID card.account password and bank card,which are poor reliability and have low security coefficient,so they are no longer satisfy-people's demand for information security.in this opportunity,biometric recognition enters people's wor-ld and are recognized by the society with its advantages of reliabilit.,convenience and practicality.As a kind of biometric recognition.iris recognition is considered to be the most reliable and effective biometric technologies because of its higher accuracy and convenience.So the biometric recognition has broad prospects for development and application market.Iris recognition system is composed of image acquisition,image preprocessing,feature extraction and matching recognition.In the system,first of all.collect the iris image with the image acquisition equipment;then.preprocess the image.which includes the image quality evaluation,image location and normalization,the image quality evaluation includes image definition,eyelid and eyelash detection and so on;Then,extract the features of the normalized iris image.and the iris features are expressed in digital form and stored by encoding:Finally,match the iris featuresxwith the feature template,and give the recognition result.In the iris image preprocessing,image quality evaluation plays an important role.On the one hand,image quality evaluation can ensure the qualification of iris image and prevent poor quality images from entering the iris recognition system;on the other hand?with the booming of the machine learning,the requirements of iris recognition system for feature extraction and matching are reduced,while the image quality is becoming more and more important.Therefore,the research of image quality evaluation is very valuable and necessary.This paper mainlv studies iris image clarity evaluation and eyelash detection.Now.most of the existing methods are time-consuming and complex.Considering the requirement of speed and accuracy in practical application,this paper proposes a new fast image clarity evaluation method.The method consists of three steps:firstly,set an adaptive threshold according to the characteristics of the reflective area in iris pupil;secondly,search the bright spot area and select the reference point;finally,select the region of interest of the iris image adaptively according to the law of image statistical analysis and evaluate the clarity of the iris image in spatial domain.Experimental results show that,the method could evaluate iris image clarity accurately with low computational complexity and high velocity,it has good application performance.Then.this paper proposes a fast eyelash detection method in iris image based on gray-scale morphology.First of all,determine the region of interest of the eyelash based on the pupil position;then.design the structural elements,and perform morphology closing operation on the region of interest of the eyelash and obtain the difference image and generate the binary image;finally,detect the eyelash from coarse to fine in the iris image according to the binary image,and the location of the eyelash are finally marked.In order to prove the effectiveness of the methods proposed in this paper,we carry out experiment on the CASIA-IrisV3-Lamp image database.The experimental results show that the proposed method can distinguish images with different clarity levels quickly and accurately,and has achieved good results in eyelash detection,which has a guiding value for practical application.