Image Encryption Based On Hopfield Chaotic Neural Network And Symmetric Segmentation Diffusion

At present,most of the deep encryption algorithms can be cracked by plaintext attack,while the encryption method using only chaos system has some problems,such as small key space,not good nonlinear characteristics,and insufficient anti-attack ability.Therefore,this paper proposes an image encryption algorithm based on Hopfield chaotic neural network and symmetric segmentation diffusion,referred to as H-BS algorithm.At first,using a hash function for clear image to generate the hash array,according to the array to determine the control parameters and the initial value of chaotic system,through the chaos system to generate pseudo random sequences,chaotic pseudo-random sequence as Hopfield neural networks initial external conditions,and through the multidimensional nature of neural networks,generated the random sequence after optimization;Then,the plaintext image is divided into N*N pieces and scrambled.After scrambling,each piece of image is scrambled in two directions.Merge the scrambled blocks into an image,which is the final scrambled result.Finally,the optimized random sequence is used to generate diffusion matrix,and the scrambled image is symmetrically segmented along its diagonal direction.Through the symmetric diffusion model designed and realized in this paper,the symmetric diffusion operation is carried out on the segmented image to obtain the final encrypted image.In this paper,several images are selected for experiments to verify the rationality and security of the algorithm,and octave simulation software is used to test the algorithm,and a series of security analysis,such as key sensitivity,histogram and correlation,is carried out.The ideal value of the information entropy of the image encrypted by the algorithm is 8,so one of the criteria to judge whether the algorithm has good performance is to calculate the information entropy.The information entropy of the image encrypted by the algorithm is very close to the ideal value,and the NPCR of the image meets the requirements,which is close to or greater than 99.6093%,and the UACI value is close to or greater than 33.4635%.Then the robustness of the algorithm is verified by simulation of clipping and noise attack.The paper has 16 pictures,9 tables,and 57 references.