Research On Preprocessing Algorithm Of Iris Recognition

As the security of information is taken into account increasingly,people have a new understanding of identity nowadays.People have higher expectations of identity security and accuracy.The traditional means of identification cannot identify effectively and automatically.In this context,identification technologies which is based on biometric produce.Iris recognition technology has become the focus of research in this field by virtue of its various characteristics.The preprocessing process of iris recognition includes image acquisition,image quality assessment,iris location,normalization and image enhancement.The main purpose of the preprocessing is to provide high-quality input for subsequent feature extraction and matching,so the results of it directly affect the performance of the entire iris recognition system.In this paper,we mainly study the preprocessing algorithm of iris recognition,and improve the accuracy and efficiency of it by deeply analyzing the image quality assessment method and iris location method.Studied in this paper mainly focuses on the following aspects:1.This paper proposes a method of iris image assessment based on BP neural network.In this paper,the quality of the iris image is systematically evaluated by three processes.Firstly,this paper detects the effectiveness of collected images.Once the images are forged,they are immediately removed.This step can greatly improve the safety of the system and avoid the waste of time unnecessarily.After that it performs the rough evaluation of the image quality.These indexes have low computational complexity and can be used to filter the images quickly and effectively.They can provide reliable input for fine evaluation of iris image quality.Then a fine evaluation of the iris image quality is performed.This paper analyzes the shortcomings of some traditional indexes,improves the calculation methods of these indexes,proposes the fine evaluation index of this paper.Finally,the BP(Back Propagation)neural network is used to fuse these evaluation factors to get the classification set.2.This paper proposes an improved iris location method.The algorithm divides the iris localization process into the process of coarse and fine positioning of iris inner circle,coarse and fine positioning of iris outer circle.For the iris inner circle positioning: First the iris image is binarized and denoised,through the projection of pupil area to perform the coarse positioning of the iris inner circle.The coarse positioning range effectively isolates the eyelids,eyelashes and other noise interference,and includes the pupil part completely.And then use the Canny operator to detect the edge of the pupil,after that this paper uses the least squares method for the inner circle boundary fitting,thus completing the iris inner circle fine positioning.For the outer circle positioning: First,according to the results of the inner circle,this paper sets up the detection template,from four directions of the pupil to detect the outer boundary point.Then we obtain the cylindrical outer positioning according to these boundary points.According to the coarse positioning parameters,this paper reduces the incremental range of the circle center and the radius of Daugman 's template,which greatly reduces the amount of calculation,finally getting the fine positioning of the outer circle.3.The overall algorithm of the preprocessing of this paper is verified experimentally.In this paper,the experimental results show that the preprocessing algorithm has a high correct recognition rate and improves the performance of the iris recognition system,which is based on the overall performance of the iris recognition system.At the same time,the quality evaluation method and the iris location method are experimented separately,which shows the advantages of the two methods compared with their similar algorithms.In summary,the algorithm of iris recognition preprocessing is studied in this paper.This paper proposes a method of iris image quality assessment based on BP neural network and a method of iris location based on small-scale search,least squares curve fitting and Daugman 's circular template detection.