Research And Implementation Of Key Techniques For Finger Vein Identification

With the development of computer network technology and the development of social information,information security has been paid more attention.As a new type of biometrics,veins are used as a basis for the extraction of subcutaneous veins,including the essential requirements of biological characteristics such as uniqueness,universality and stability.Furthermore,it has the characteristics of living certification,non-contact acquisition and so on.what is more,it can overcome the shortcomings of mature features such as fingerprints.Compared with the palm vein and the dorsal vein,the finger vein has the advantages of small collection device and good user friendliness,and has gradually got the attention of the scholar.Because of the short development period,the technology needs to be improved in image preprocessing,image vein texture extraction and vein feature matching to meet the actual needs of the product.In this paper,through the analysis of the technical principles of finger vein recognition,the research work on the above aspects is as follows.(1)Designed a finger vein collection program.Aiming at the problem of hardware equipment for finger vein collection,two popular acquisition schemes were analyzed and studied.According to the existing hardware conditions and the collection environment,the vein collection device which can meet the experimental requirements of this paper is completed and using the device to establish a finger vein image library.(2)A method of finger angle correction is proposed,and the method of ROI region extraction based on bidirectional gray scale projection is improved.The image in the collection process will have an indefinite tilt of the finger.According to the shape of the finger itself,this paper presents a method based on finger center centroid fitting center line to complete the angle correction.On the basis of image angle correction,the method of intercepting ROI region by traditional bidirectional gray scale projection is improved,and the effect of ambient light on finger vein area is better overcome.Through the experimental verification,the preprocessing algorithm in this paper has achieved good results in image angle correction and ROI extraction.(3)A finger vein extraction algorithm is proposed.This paper analyzes the basic principles of finger vein extraction,and focuses on two kinds of vein pattern extraction algorithm.Aiming at the problem of the quality of vein texture extraction and the efficiency of the algorithm,an algorithm for extracting vein lines based on multi-directional gradient coding is proposed.The experimental results show that the algorithm has great advantages in vein texture extraction quality and algorithm efficiency.This algorithm can also have better performance on low-quality vein images.(4)The idea of fuzzy decision is merged into feature extraction and matching algorithm based on global gray normalization.This paper discusses the categories of feature extraction and matching algorithms and analyzes their strengths and weaknesses separately.Based on the study of three classical feature extraction and matching algorithms,the feature extraction and matching algorithm of global gray scale normalized product is improved by combining the idea of fuzzy decision.The experimental results show that the proposed algorithm is better than the other three algorithms in terms of accuracy and algorithm efficiency.(5)Design and implement a finger vein recognition simulation framework at the PC.This paper introduces the structure and operation flow of the framework.The framework can perform functional demonstrations(registration,matching)of finger vein recognition,as well as the results of algorithmic experiments in each intermediate link.