Abstract
More recently, the exponential growth of the Internet hascreated the potential for both business-to-business transactionsand business-to-consumer transactions. It is desired to conductthese transactions across open networks. It is inevitable toconstruct a security system for computer networks in order todiscard the effects of attackers and such that no person can do an unauthorized transaction. In order to keep the security ofcomputer systems, user authentication is a necessary step. Afterfinishing the user authentication, the messages are encryptedto send in communication.Biometric encryption is the process of using a characteristic of the body, such as the pattern in a fingerprint, to code orscramble data. Since each finger pattern is unique, the way inwhich the data is scrambled is also unique. The public-keycryptosystems can establish authentication and provide digitalsignature to assure the authorship and authenticity of data. Hence we propose an authenticated encryption system based onthree-factor authentication: smart card, fingerprintauthentication, and public-key cryptosystem. The system usesminutiae points as a template stored in the smart cardchip(fingerprint-card) to identify users and then further uses the minutiae points as the private key to sign, encrypt, and decrypt electronic documents. It is a digital signature scheme integrated with the public-key cryptosystem and fingerprint authentication, so it provides secrecy, authenticity, nonrepudiation, and integrity, simultaneously.It is necessary to develop a good matching algorithm with a higher speed and a lower false rejection rate(FRR) for fingerprint authentication. We have proposed a fast point matching algorithm which is invariant to translations, rotations, scale changes, local distortion, and extra/missing points conditions. The proposed algorithm is fast and provides lower false rejection rates for fingerprint authentication. The computational time is less than 0.32 s for 50 to 50 point matching on a Sun-Sparc4 workstation. For matching 800 fingerprint images, if the false acceptance rate is set at 0.1%, the false rejection rate is at 1.03%.To test the performance of our proposed point matching algorithm, we apply the algorithm to automatic seal identification. The identification system uses the point matching algorithm to register the seals quickly and accurately. To increase the identification rate, a fake detection algorithm is proposed to remove the over-ink areas and detect fake areas in the registered images. Experiments prove the identification ability for the proposed approach. The processing speed for identification is less than 2 seconds on a Sun-4 workstation.Public-key systems tend to be significantly slower than secret-key systems. They are applied to a small amount of data. The Private-key(symmetric) systems have high data rates and low power consumption features. Hence it is necessary to use private-key encryption algorithms for high data rates communication. DES(Data Encryption Standard) is a U.S. and international standard for private-key encryption. Today, the cryptographic community is quite aware of the fact that breaking DES is getting more feasible than it was previously thought. Hence we propose a new secure block cipher SEA2(Secure Encryption Algorithm 2) for use in our authenticated encryption system. Our analysis shows that attacks based on differential cryptanalysis and linear cryptanalysis do not offer any significant improvement over a brute force attack for 4-round SEA2. The C language implementation on 4-round SEA2 could achieve the speed of about 22.7 Mbits/sec on a 200MHZ Pentium PC, running Windows 95. It is a fast and secure encryption algorithm.