Research On Palm Vein Recognition Algorithms

With the development of science and technology,more and more people pay attention to the security problem.How to accurately and reliably identify a person's identity has become an urgent problem to be solved.The traditional identity recognition mechanism has the disadvantages of easy to steal and copy,which can not meet the requirements of high security.Biometrics is an effective way to solve this problem.Among them,palm vein recognition technology is a kind of biometrics recognition technology in recent years,which has the advantages of high anti-counterfeiting,high recognition precision and easy acceptance by users.In this paper,we mainly study the comparison of ROI(Region of Interest)images and the image enhancement and denoising methods of ROI images.This paper focuses on the extraction and matching method of NBP(Neighbor Based Binary Pattern)feature,the extraction and matching method of SIFT(Scale Invariant Feature Transform)feature,and proposes a fusion texture feature and local invariant feature of the palm vein recognition method.The main work and research results of this paper are as follows:1.In this paper,the preliminary study and comparison test of the acquisition and pretreatment of ROI images of palm vein recognition were carried out.We selected the ROI image extraction method based on the palm rectangle and CLAHE(Contrast Limited Adaptive histogram equalization)method and the median filter method to image enhancement and denoising to achieve the best of the subsequent test.2.In this paper,two kinds of single-feature palm vein recognition methods based on NBP feature and SIFT feature are studied.Aiming at the drawback of traditional RANSAC(Random Sample Consensus)method in SIFT matching process,a method based on similarity distance deduction error matching point is proposed to improve the efficiency and accuracy of error matching point elimination,so that SIFT algorithm in the palm vein recognition in the application of more accurate.Then,the design experiment is used to analyze the two kinds of palm vein recognition methods,and the performance characteristics of the two algorithms are obtained,which provide the basis and ideas for the following two algorithms.3.In view of the advantages and disadvantages of the two algorithms studied earlier,the feasibility of the fusion of the two algorithms is analyzed: NBP feature as a global feature,SIFT as a local feature,both for different palm image discrimination and the same palm image matching degree has a strong complementarity.The NBP feature can make up the SIFT feature without real-time disadvantage in the experimental time.The robustness of the SIFT feature can compensate for the poor robustness of the NBP feature on the larger displacement.We get the conclusion that the two algorithms are very suitable for information fusion,a palm vein recognition algorithm based on texture feature and local invariant feature is proposed,which combines the two algorithms effectively.The correct recognition rate of the algorithm is improved,and the correct recognition rate is 99.114% and 99.722% respectively in the PolyU palmprint and the laboratory palmprint.