New Fuzzy Vault Algorithm Based On Fingerprint

With the frequency and universality of people communicating with each other on the internet,people have paid more and more attention to the internet security. Information security problem hasbecome a key technical problem in network environment, especially the cryptography plays a moreand more important role in the network information security field. In the cryptography, the keysecurity is very critical. Sometimes, the key is so long that it is difficult for the users to remember,so they use a password to encrypt and protect the key, however, it is vulnerable to dictionary poorguessing attacks, and security is very low. Because of the traditional password encryptionalgorithm’s weaknesses, biometric encryption is proposed to protect the key. It combines thebiometric characteristic with the key in a certain way，it eliminates a lot of trouble of storing andmanaging the key, and the key itself can encrypt and protect the biometric, which increases thedifficulty of the attacker.In the biometric encryption, fingerprint is most widely used and it has the best developmentprospect. Because the fuzzy vault algorithms can resolve the contradiction which between theambiguity of biometric characteristic and the accuracy of the key, so this paper mainly studies thenew fuzzy vault algorithms based on the fingerprint features and it contains the following severalparts:Firstly, this paper deeply studies and analyzes the fingerprint identification technology, it doessome researches on the fingerprint image preprocessing techniques, with comparing and analyzingseveral different fingerprint feature extraction algorithms, it studies the minutiae-based fingerprintfeature extraction algorithms, and makes a complete simulation. It has laid a good foundation forthe fingerprint key generation.Secondly, in order to improve the key recovery rate, this paper proposes a new fuzzy vaultalgorithm based on fingerprint and Chinese Remainder Theorem. In the process of encryption, itintroduces the positive integers which are pair-wise prime, password key is combined with thefingerprint features and the positive integers can be used to generate fingerprint key, so an attackermust have the positive integers which in the encryption process and the correct fingerprint featurepoints that can break the fuzzy vault, then get the key, it has greatly increased the difficulty to theattacker. In the process of decryption, the key can be recovered by using the Chinese RemainderTheorem according to the threshold idea. According to the theoretical analysis and experimentalresults, the key’s safety and the rate of recovery are increased.Once more, this paper proposes a new fingerprint fuzzy vault algorithm based on double polynomials for reducing the legitimate user’s computational complexity, the algorithm takesadvantage of the verification data to check the validity of elements in the unlocking set, it is mainlyup to removing the noise pointes before the polynomial reconstruction, so the valid user’scomputational complexity can be greatly decreased. Experimental results show that the legitimateusers’computational complexities have been significantly reduced.