Strengthening Cryptographic Keys Using User’s Voice Biometric

Nowadays, the most trustworthy tool to protect our information is based on cryptography.To enhance cryptography, it is highly important to select a robust cryptographic key that implies a high degree of uniqueness and randomness.In general, users lack proficiency in generating strong cryptographic keys.Furthermore, they often struggle with remembering how to generate strong cryptographic keys.This paper introduces an interesting method to abridge and boost the key generation step.The proposed method makes use of audio files that contain the user's voice biometric.A variable histogram-thresholding, which is calculated for each audio file distinctly, was used to get approximation peaks in the audio file and then to extract the discriminating features that ensure the uniqueness.The features will then be normalized and converted from a decimal to binary system to generate two 128-bit subkeys.Finally, to generate a robust AES-128-bit encryption key, these two subkeys are shifted, mixed, and XORed to enhance randomness.A dataset of more than 900 audio files is used.Testing and evaluation demonstrate how robust the suggested system is for helping users generate robust cryptographic keys.