The Retinal Identification Based On Principal Component Analysis

Objective: With the increasing need for identity recognition,the traditional identification technology and fingerprint based identification technology can no longer meet the needs of people,and the retina,a biometric feature,has entered people's attention with its high security and other advantages.The purpose of this paper is to study and design a retina recognition algorithm which is not affected by the angle of input image,scale transformation,illumination inequality,and can realize real-time matching at the same time.The identification system based on retina recognition algorithm is more difficult to cheat than iris,face recognition and other systems,and more suitable for military or government departments with high security requirements,so the study of this paper is of great significance.Research data: The data used in this paper is from the drive international open database.There are 20 training set data and 80 test data in total,which are obtained from the rotation and scaling of color fundus images in the test set.Research methods: In the aspect of system design,firstly,Hough transform,bilinear interpolation algorithm and angle rotation are used to preprocess the image,and then the fused Gaussian matched filter and two-dimensional maximum entropy threshold segmentation method are used to obtain the binary image of retinal vessels.Principal component analysis is divided into two parts,one is to read and input the template image of the data to be matched.Here,the binary vascular map is selected as the object to be processed.After comparative study,the cumulative contribution rate is set to 85%,and the dimensionality reduction and projection of the image in the database to the feature space are completed.The other part is to reduce the dimension of the input image,then measure the similarity between the input image and the image in the database by Euclidean distance,and select the image with the shortest distance as the matching image.If the output image is consistent with the input image,the matching is considered successful.Results: The recognition rate of this algorithm reaches 96.25%,which is higher than other algorithms.The image after rotation and scaling can still be matched successfully.Because the algorithm also carries out the blood vessel segmentation,so the running speed is slower than the existing algorithm,and the time to recognize an image is 2.6012 seconds.In the part of retinal vascular segmentation,the accuracy and specificity of this paper are the same as other algorithms,but the sensitivity is 0.8,which is higher than the traditional algorithm.