The Fuzzy Vault Scheme Based On Multi-biometric

Biological characteristics have been widely used in information encryption and key protection domain because of its stability,uniqueness and portability.Fuzzy vault is a popular key protection scheme,but the classic fuzzy vault only uses a single biometric template.As people have higher request for recognition rate,template security and universality,the fuzzy vault with only one kind of biometric template already can't meet the demand.At the same time,the multi-biometric fuzzy vault scheme can effectively improve the recognition rate,reliability and security,thus some defects of the single biometric fuzzy vault scheme can be overcome,so it is now to be favored by people.However multi-biometrics also brings an increase in the number of templates.How to effectively use the multi-biometric templates and protect the multi-biometric templates is the key problem that the multi-biometric fuzzy vault scheme must be faced.At present,the multi-biometric fuzzy vault scheme is still in a development stage,and it is a very worthwhile research field.In this thesis,the following aspects of the work are done in the field of the multi-biometric fuzzy vault scheme:(1)In this thesis,we firstly introduce the implementation of biometric image preprocessing and the feature extraction in the fingerprint biometrics technology,finger vein biometrics technology,finger crease biometrics technology,and study the classic fuzzy vault scheme while analysis its locking and unlocking process in detail.(2)In this thesis,based on the analysis of advantages and disadvantages of the multi-biometric fusion technology and the feature fusion level,the fuzzy vault scheme based on the fingerprint and finger vein feature level fusion is proposed.The fingerprint and finger vein overlap feature template is used to increase the number of real feature,and we encode per feature point in grid manner in order to obtain a fusion binary bit string that can hide the genuine feature point information.Experimental results demonstrate that this scheme can provide better reliability and the security of template.(3)In this thesis,on the basis of multi-biometrics data level fusion,the fuzzy vault based on the data level fusion of finger vein and finger crease is proposed.A new fusion template is formed by the combination of finger vein preprocessing image and finger crease preprocessing image in data fusion level.Due to that the fusion template not only includes new features information but also includes the original feature information generated from finger vein and finger crease,so we introduce the template conversion technology.Experimental results demonstrate that the proposed scheme has a significant improvement in reliability compared with the single biometric fuzzy vault and improves the security of biometric template by introducing the template conversion technology.