With the increasing reliance on electronic information, which needs to be exchanged across the internet or stored on open networks, cryptography is becoming an increasingly important feature of computer security. A biometric key dependent cryptosystem is proposed, to ensure the security of the whole system by using iris pattern as a key in a cryptosystem, like, Key-dependent Advanced Encryption Standard (KAES). KAES is used to ensure that no trapdoor is present in cipher and to expand the key-space to slow down attacks. The proposed system gave significant results under various tests for the key uniqueness and the system randomness.

Face recognition could be applied to a variety of practical applications and problems, including security and criminal identification systems. Face recognition using eigenface approach was motivated by information theory as it provides a practical solution. In this paper, the Principal Component Analysis (PCA) is used for eigenfaces (eigenvectors) computation. These eigenfaces present the extracted features for the faces to be recognized. A multilayer Artificial Neural Network (ANN) with back propagation adaptive learning algorithm is used for the classification phase. A number of experiments have been conducted on the system using the Olivetti Research Laboratory (ORL) database. Promising results have been achieved. Total performance accuracy on the data set used reached 98%.