Key Generation Based on Facial Biometrics

Over the past few years, advances in communication technology have brought large amounts of digital data to ordinary media, which required the development of computer security technologies. In this work, we suggested a new method to generate a biometric key to encrypt data using the properties of the human face, then used this key to encrypting speech messages and hide them inside the colored images. This can be achieved depending on splitting the facial image into two parts (upper and lower parts) and then generated a unique encryption key using Maximum-Relevance Minimum Redundancy (mRMR) feature selection algorithm from the upper part after that encrypted the original speech message using two levels, in the first level we used Arnold cat map to permutation the samples then in the second level used bio-key to encrypting the message and then hide the encrypted speech message in the lower part of the facial image. In order to determine the efficiency of the proposed method, different measures were applied (correlation coefficient, PSNR, MSR, SSIM).