Cryptography based on Fingerprint Bio Metrics

Abstract

Fingerprints remain consistent and stable throughout a human’s lifetime. This research exhibits the utilization of fingerprint bio-metrics to generate secured keys for improved security. The main contribution is the generation of 87 keys, achieved by enhancing the fingerprint image and sharpening it with a Laplacian filter. Three types of geometrical shapes chosen arbitrary, such as circle, square, and triangle, are separately drawn on the fingerprint after binarization, morphological operations, and thinning. Each shape is drawn five times on the fingerprint with different radii for each type to increase the number of keys. The end and bifurcation points are extracted as features inside and outside these shapes and these features are considered as keys. Chaotic-Pseudo-Random-Number Generator (CPRNG) technique is used, and the generated keys are merged with those generated from the positions of minutiae (end and bifurcation) points. The process was implemented using MATLAB R2021b. The simulation results demonstrate that it is difficult to crack the keys generated by this technique because the attacker requires a very long time, almost 7.5595e+159 years, to decrypt the encrypted message. Using geometrical shapes and CPRNG) technique increases the number of keys. The contribution of 87 keys is raising the time needed to break the encrypted text to this time, which is greater than the time required to crack the keys generated in some previous research compared to them. Therefore, the proposed technique enhances privacy and security. It can be used via deep learning in fingerprint identification, recognition, and key generation mechanisms and can use any other geometrical shapes.