Normalized Chaos-Based Nonlinear Image Encryption Using Biometric Key

This paper proposes a multi-biometric encryption system that harnesses the strengths of both face and speech biometric data, ensuring a highly secure encryption framework. Our method incorporates an enhanced nonlinear sigmoid function, ensuring the encrypted data remains resilient to cryptographic attacks. To encrypt input face images, we adopt a dual-layered permutation-diffusion preprocessing architecture to pass by a modified sigmoid function, which effectively scrambles and cyphers the face image data into encrypted templates. Additionally, the speech templates are used to extract biometric features through SVD (singular value decomposition), serving as a diffusion key. In contrast, the other keys are generated from a quadratic chaotic map. We conducted a comprehensive performance analysis such as correlation rate, peak signal-tonoise ratio (PSNR), and mean square error (MSE), were utilized to evaluate the effectiveness of our proposed algorithm. Simulation results demonstrate the high performance and robustness of our scheme against cryptographic attacks, outperforming similar existing approaches.