Enhancing Robustness of Image Encryption Algorithms with Adaptive Approach

To enhance the robustness of image encryption algorithms, we propose a new adaptive image encryption algorithm. The implementation begins by dividing grayscale or color images of any size into 2×2 sub-blocks. In the code implementation, this can be achieved using block processing functions like mat2cell() in MATLAB or similar image partitioning methods in Python. We then utilize the pixel intensity values from the top-left sub-block combined with Chebyshev chaotic mapping to generate a matrix of equal size to the top-right sub-block. The chaotic mapping implementation typically involves iterative calculations using recurrence relations x_{n+1} = cos(k * cos^{-1}(x_n)) to generate pseudo-random sequences.

This generated matrix is applied to replace the pixel values in the top-right sub-block through element-wise operations or XOR transformations. The encryption process continues clockwise through all sub-blocks using similar matrix-based substitutions until the top-left sub-block completes encryption. The algorithm's design incorporates a pixel value diffusion mechanism through sequential sub-block processing, where each encrypted block influences the next, creating strong diffusion and confusion characteristics. This approach enhances cryptographic security through dependencies between adjacent blocks and chaotic sequence integration, making it suitable for various image types without size constraints while significantly improving encryption complexity.