Image encryption algorithm based on improved Thorp and local variance adaptive diffusion

Abstract Natural images with complex textures and detailed information are a major challenge in the field of image security, and existing encryption methods are usually difficult to deal with the fine details of such non-uniform images, for this reason, an image encryption algorithm based on Thorp permutation with local variance hierarchical diffusion is proposed. Firstly, the chaotic map 2D-CIM is constructed with a larger key space. Secondly, Thorp permutation based on Julia set fractal sequence is used to scramble, which effectively reduces the correlation between pixels. In the diffusion stage, adaptive diffusion based on local variance is applied to enhance the encryption process by adjusting local image features. A large number of experiments show that the algorithm has strong anti-attack ability, and it is difficult for the attacker to recover the details of the plaintext image by analysing the local features of the image, which is suitable for high security application scenarios.