Nonlinear biometric pre-processing applied to image encryption scheme

Abstract

In this paper, we propose an efficient preprocessing approach for an image encryption scheme based on a nonlinear function, namely the sigmoid function, applied to a permutation diffusion architecture. We introduce several chaotic functions and an original image fingerprint, which jointly form the encryption key of the proposed scheme. To evaluate the effectiveness of our algorithm, we use performance metrics such as mean square error (MSE), peak signal-to-noise ratio (PSNR), and correlation rate. Our results demonstrate the robustness of the proposed encryption scheme. In addition, the scheme is shown to be more efficient and superior to existing encryption schemes in the literature through cryptographic attacks.