Fuzzy Vault Key Algorithm Based On Fingerprint Feature

With the development of computer and communication technology, information security gets more and more attention. Cryptography is the foundation of information security and the basic premise for modern information network to exist and develop. The security of key is a crucial problem of cryptography, but the key management of traditional cryptography has some defects. For example, the traditional key is a pseudo-random number with a certain length, and the greater the randomness, the more secure the system, but the more difficult to remember the key. Therefore, many people use a short password to encrypt the key. Then the system's security just depends on the security of the short password, and it is vulnerable to dictionary attacks and very unsafe. Key algorithm at the basis of biometric characteristic could solve the key management problem of conventional cryptography. It combines the biometric characteristic with the key according to certain rules so as to increase the safety of the system. There is great conflict between the vagueness of biometric characteristic and the accuracy of the key cryptography, and fuzzy vault algorithm can resolve this conflict, so it becomes a typical application for bio-key algorithm.Fingerprint is the leading biometric characteristic in term of application, so it has a vast development prospect. This paper mainly studies the fuzzy vault key algorithms based on fingerprint feature. It contains the following three parts:1. It analyzes and discusses fingerprint identification technology. It does a series of in-depth researches on the fingerprint image acquisition and the fingerprint preprocessing, such as image segmentation, normalization, direction computation, filtering enhancement, binarization, thinning and so on, and makes a complete simulation. Then we do the analysis of fingerprint feature extraction algorithm on the preprocessed image. Finally, a fingerprint matching analysis is done.2. We propose a fuzzy vault key encapsulation algorithm based on password. At the basis of the fuzzy vault algorithm, we first encrypt fingerprint minutiae with the password, then apply the encrypted points in the fuzzy vault, so that the attacker must have both the fingerprint feature template and the correct password to break the vault to get the key. Simulation result shows that: compared to the fuzzy vault scheme by Nandakumar, locked vault using the fingerprint minutiae,which is encrypted by the password, significantly reduce the system's false acceptance rate (FAR), and improve the genuine acceptance rate (GAR) to a certain extent. So the system security can be improved.3. Aiming at the problem of low complexity of the fuzzy vault based on the minutiae, a new fuzzy vault key algorithm based on the fingerprint feature vector is proposed. The extracted fingerprint image is divided into sectors taking the core as the reference point and normalized, and then the fingerprint feature vector is obtained by the application of the filter. Finally, the fuzzy vault is constructed based on the fingerprint feature vector. Our simulation shows that, the attack duration of the algorithm is much longer than the fuzzy vault key algorithm based on the minutiae, and also the FAR/GAR performance is ideal.