Generation of Encryption Keys Related to Encryption Algorithm Using The Most Important Features of Biometrics

Abstract

    Cryptographic algorithms are under regular attacks due to poor cryptographic keys. The research introduces two new algorithms for generating cryptographic keys based on biometric descriptions developed from an original algorithm. The algorithms rely on fingerprint images and AES encryption algorithm. The research suggests extracting fingerprint features using the mintuae method and then generating cryptographic keys based on these features. Key generation operations include several steps that increase cipher strength such as switching and mixing operations, fingerprint image problems were avoided through pre-processing operations. The Extracted features from images were independent of shape, direction and rotation. The two modified algorithms were able to successfully encode and decode the texts with less implementation times than the original algorithm and with higher performance. They also increased the sobriety of the algorithm towards attacks as the algorithms were tested on fingerprint images of different people and show the algorithm's ability to address it. They prevented the decryption of the original text in case the fingerprint is not correct. The modified first algorithm outperformed the second modified one and the original algorithm in terms of time, performance, and attack. It also managed to decode the texts using inverted, rounded, partial and left-hand fingerprint images.