Adaptive Block-Based Color Image Encryption Using Sine and Logistic Chaotic Maps

Abstract

This paper presents a robust and reproducible chaos-based encryption scheme for color images by combining the Sine map and a double-stage Logistic map within a block-based framework. Unlike conventional approaches that rely on fixed parameters or unclear key derivation mechanisms, the proposed method dynamically generates chaotic parameters from the statistical properties of the plaintext image, ensuring strong key sensitivity and resistance to cryptanalytic attacks. Initially, the input RGB image is decomposed into three color channels and divided into non-overlapping blocks of size 16$\times$16. In the first stage, a Sine map is employed to generate chaotic sequences for each color channel using a control parameter derived from the mean variance of the RGB components. These sequences are generated without direct modification of the image, establishing an initial confusion layer. In the second stage, a Logistic map is used to produce a unique permutation of block indices, enforcing strict uniqueness and enabling a selective XOR diffusion process between image blocks and the previously generated Sine sequences. This operation produces an intermediate encrypted image. To further enhance security, a second diffusion stage is introduced. New Sine map sequences are regenerated using the same chaotic control parameter but with initial conditions dynamically extracted from the intermediate image, ensuring plaintext dependency. A second Logistic map with distinct parameters is then applied to generate a new unique block index vector, followed by an additional XOR operation, yielding the final encrypted image. Comprehensive experimental analyses demonstrate that the proposed scheme achieves near-ideal entropy values, negligible pixel correlation in all directions, and strong resistance to statistical and differential attacks. The decryption process accurately reconstructs the original image when the correct keys are provided, confirming the reversibility and reliability of the method. Owing to its simplicity, high security, and computational efficiency, the proposed encryption framework is well suited for secure image transmission and storage applications.