The Research Of Biometrics Fusion Encryption Method Based On Fuzzy Vault

Fuzzy vault is an encryption framework to protect the user’s identity key using the biometric. This framework is of a great balance between the fuzzy biometric and the accuracy of an encryption algorithm, while it is of a high safety. So we can see it has caused a wide range of research interests of the researchers all around the world. Because of the defect susceptible to theft and forgery of single biometric identification system, and the inevitable interference and error of the biometric template when building, we’ve work out a method to fuse multi biometric and encrypt the key of users based on fuzzy vault framework. This method can deal with the template forged by the traditional biometric encryption method, cross-matching and other hidden dangers and effectively enhance the encrypted biometric authentication effectiveness and robustness.The main works and the innovation points of this paper are:1. We have done a thorough and meticulous research on the traditional biometric identification technology and the emerging biometric encryption technology in recent years. We introduce the key technology and the development of them. And then, we introduce the main methods of biometric encryption all around the world.2. Based on extensive research and preparatory work we’ve done in related fields, we worked out a more effective and reliable multi-mode biometrics fusion encryption method than single source biometric encryption. This biometric encryption scheme can be divided into two levels:feature level and key level. In the feature level, we do space affine transformations and the fusion operation for the feature templates using the abstract of user’s key, and get the fusion cross-correlation feature template. Then we do the encrypting and decrypting operations for the fusion biometric template according to fuzzy vault scheme. The advantage of feature level fusion is that we can protect the users’biometric templates and then reduce the harm of the disclosure of one’s unique biometric templates. In the key level, we divide the user’s key into some parts according to the scale of each mode biometric feature, and then encrypt the parts of user’s key using a random sequence and get the sub keys. For each sub key, we do the encrypting and decrypting operations according to fuzzy vault scheme. After these, we recover the parts of original user’s key using the auxiliary random sequence and combine them and get the user’s entire key. While, in this paper, we still worked out a joint validation process that can further improve the effect of template protection and identification of this method.3. For the dual-mode features of different fingerprint from the same source, we give an implementation scheme of this multi-biometric fusion encryption method and test it using the multi-biometric database from the same source made by the research team of Shandong University. And finally we find it is more effective with a higher succeeding rate and can protect the information of biometrics and keys better.