Font Size: a A A

Design And Implementation Of An Attendance System Based On Finger Vein Recognition Technology

Posted on:2021-02-19Degree:MasterType:Thesis
Country:ChinaCandidate:D D ZhaoFull Text:PDF
GTID:2438330602997840Subject:Control Engineering
Abstract/Summary:PDF Full Text Request
Identification technology occupies a very important position in personal information security.With the advancement of technology and the improvement of people's living standards,more and more people attach great importance to personal information security.Traditional identification is to use external media such as keys and passwords for confirmation.This type of identification method is very inconvenient and easy to lose and be stolen.As a new identity verification technology,the main principle of biometric identification technology is to identify differences in specific positions in the human body,such as fingerprint,face and finger vein.Biometric identification technology has developed rapidly because of its security and convenience.Finger vein recognition technology is one of the important biometric recognition technologies.Therefore,the research on finger vein recognition technology in this paper mainly includes the following points:(1)During image acquisition,different poses of user's finger result in the offset and rotation of finger.In view of the above problems,a method based on edge detection for extracting the region of interest of finger vein is proposed.This method first extract the edge of finger using canny operator,split out the finger area and calculates the center curve of the finger.Then fits he center curve of the finger using least square method,calculates the deflection angle and rotate the original image in responding angle.At last,the histogram statistics of pixel gray values are used to determine the reference point and complete the extraction of the region of interest.This method can overcome the problem of finger offset and rotation,and ensure that the region of interest extracted is the same region of the finger(2)Aiming at the problem of unclear finger vein in the image,A method based on spatial filtering fusion is proposed to enhance the image.For darker images,the method first uses Gamma algorithm to improve the overall brightness of image;for the problem of low contrast,CLAHE was proposed to enhance contrast;at last,bilateral filtering was introduced to filter the image noise while keeping edge information.Experiments show that this method can improve the contrast and entropy information to a certain extent,and improve the recognition rate,especially for poor quality images.(3)In order to improve the accuracy of finger vein recognition,a center loss function is introduced.The loss function can adjust the feature vector of each category,so it can minimize the distance of intra-class and maximize the distance of inter-class,and then greatly improve the recognition rate of the model and reduce the equal error rate of the model.In order to reduce the risk of overfitting of the model,a new regularization method is proposed,which can adjust the regularization coefficient automatically.This method can accelerate the convergence speed of the model to some extent and generalize the model.(4)In view of the above algorithm research,a finger vein attendance system is designed.This system can record the user's punch time and judge whether the punch is valid,and the user can query his relevant punch information through his account.
Keywords/Search Tags:Finger vein recognition, region of interest extraction, image enhancement, center loss, dynamic regularization
PDF Full Text Request
Related items