Research On Finger Vein Recognition Algorithm And Its Cipher Application

With the continuous development of information technology in recent years,information security has been paid more and more attention.As an important way to ensure information security,biometric identification technology plays an irreplaceable role.Compared with systems based on other biometric identifications,such as fingerprint,iris and face,finger vein recognition is a new type of biometric technology,and its main characteristic lies in using the pattern of veins within fingers as feature for authentication.Besides,finger vein recognition also has the following characteristics: vivo-identification,high-security and non-contact.Consequently,it has become a new research hotspot,and some related research about finger vein recognition have important practical significance.Firstly,this thesis introduces the basic content of the finger vein recognition system and some classical algorithms;Secondly,in consideration of the problem of finger rotation when capturing finger vein images and system security,a new method of finger vein recognition based on Radon transform and chaotic encryption is proposed.In addition,the fuzzy vault algorithm is studied,and a fuzzy vault scheme based on the finger vein feature is proposed.The main work is as follows:1.Some related processing technologies about the finger vein image are introduced,including a series of pre-processing operations,image enhancement,segmentation algorithms for the finger-vein image and some performance analysis of biometric identification system.2.A new recognition method for finger veins based on Radon transform and chaotic encryption is proposed.In this method,the finger vein skeleton features are first extracted after image enhancement,Niblack segmentation,filling and refinement,then the Radon transform is used to calculate the basic direction of the finger vein,and the skeleton features image is rotated according to the calculated direction.When registering,Logistic chaotic sequences are generated by the key which users input,then the chaotic sequences is applied to encrypt the rotated image and after tahat the encrypted images are stored in the finger vein feature database.When matching,the user need to collect the finger vein and enter the key,the features of the collected finger vein were obtained after extraction,and using the key decrypt the feature in the finger vein feature database,then by evaluating the similarity between these two features matrices,the finger vein is classified.Experimental results show that: this algorithm can get a higher recognition accuracy without a lot of template training,as well as have higher security.3.A new fuzzy vault algorithm based on finger vein feature is proposed.The algorithmcontains two parts: key binding and key recovery.In the part of key binding,the new minutiae extracting method is used to extract the minutiae of finger vein,and a new parameter "length" is defined for the minutiae.After that,a large number of chaff points are generated randomly,and each point(including the real points and chaff points)are stored in the fuzzy vault after data processing and transformation,among this,real points are blind by the key polynomial and the chaff points are caculated by radon numbers.In the part of key recovery,the minutiae coordinates of the query finger vein and the fuzzy vault are compared,then by selecting some of these matching minutiae do the opposite calculation,then the polynomial is got and the protected key is recovered.Through experiments and analysis,the results show that this algorithm is of good usability and security.