Font Size: a A A

Research On Key Technology Of Finger Vein Recognition

Posted on:2018-03-31Degree:MasterType:Thesis
Country:ChinaCandidate:W W YangFull Text:PDF
GTID:2348330542956730Subject:Electronic and communication engineering
Abstract/Summary:PDF Full Text Request
In recent years,with the development of Internet technology,information security is becoming more and more important.There is an urgent need for a more secure,more confidential,and more convenient biometric approach to secure information.Medical studies have shown that the shape of the finger vein is unique and stable.Compared with many biometric features,finger vein recognition technology has its unique advantages.First,finger vein recognition is a living body recognition,which exists in the living body,and It is not easy to be stolen and copied;secondly,it is not affected by age and finger surface dirt.Therefore,finger vein recognition has a wide range of research prospects.This paper first introduces the background and significance of the research,summarizes the development status of finger vein recognition technology at home and abroad.Then,the key technologies such as finger vein image acquisition and preprocess finger vein feature extraction and finger vein recognition are studied and explored deeply.Based on the establishment of finger vein image database,this paper proposes a regional location algorithm for finger vein image based on tangent line,which is based on the problem of finger tilt in the actual collection of finger vein image.After the edge detection,the algorithm first obtains the complete finger boundary image through the de-connected domain method,and then finds the midpoint coordinates of the straight line of the first and the last column of the finger boundary image,Vertical direction of the straight line angle as the finger vein tilt angle,and finally tilt the correction.And then the two finger boundaries to do the tangent line,cut to the finger vein image area of interest,and then use the sliding window mechanism to find the joints of the fingers,and finally the height of the two joints as the finger vein area of interest.The experimental results show that the algorithm can accurately locate the region of interest of finger vein image,and has good robustness,which lays the foundation for the subsequent image feature extraction and accurate identification.By analyzing the feature extraction algorithm based on LBP and PCA based finger vein image,a feature extraction algorithm of finger vein image based on block LBP and block PC A is proposed for the single feature description.Firstly,the ROI region is divided into several blocks,and the LBP eigenvalues of each region are extracted.Then the eigenvalues of all regions are superimposed to form the eigenvector.Then,the whole sample feature matrix is divided into different regions.Finally,PCA Dimensionality,the dimension matrix after dimensionality reduction.The experimental results show that the proposed algorithm highlights the details of the finger vein image,and greatly reduces the feature redundancy,improve the recognition accuracy and recognition speed.Aiming at the problem that the linear classifier recognition rate is not high,this paper studies and implements a finger vein recognition algorithm based on neural network.The algorithm uses the reduced dimension feature matrix obtained by the feature extraction step as the input of the neural network to train the neural network to obtain a neural network classifier which satisfies the real-time and precision requirements.Experimental results show that the algorithm has faster recognition speed and higher recognition rate than the nearest neighbor classifier.Finally,the development of a finger vein recognition software system,elaborated on the overall design of the software system ideas and the function of each module and its implementation.Through the development of the actual system,this paper provides the idea for the research of this paper,and provides the experimental platform for the practical application of the algorithm.
Keywords/Search Tags:Finger Vein Recognition, Region Of Interest Locating, Feature Extraction, Neural Network
PDF Full Text Request
Related items