Sudoku-Based Image Encryption with Iterative Thresholding and Strong Diffusion

In the current internet era, the number of information security vulnerabilities has increased dramatically. Image and text encryption have become critical preprocessing steps in secure information transmission. Sensitive information can be transmitted through encrypted images, facilitating the implementation of various secure communication systems. This paper proposes an image encryption scheme that employs Sudoku as a cryptographic key matrix, combined with a strong diffusion mechanism to enhance pixel confusion and diffusion effects. The proposed method achieves high-level pixel scrambling through multiple rounds of iterative threshold encryption, pixel padding with random shuffling, and Sudoku-based permutation. Additionally, rotation operations are applied to further increase the irreversibility of the encrypted image. The core keys include the iterative threshold sequence, row-column diffusion keys, and random premutation parameters, ensuring that the encryption is fully reproducible. Experimental results demonstrate that, while preserving reversibility, the proposed method achieves significant confusion and diffusion performance. For the Lena image, the method attains NPCR ≈ 99.22% and UACI ≈ 33.30%, indicating its effectiveness as a robust image encryption approach.