Application Of MMT Permutation And BP Neural Network In Image Encryption

With the development of science and technology,people began to frequently transmit information on the Internet.Images,as the carrier of information transmission,are also the most commonly used method of information transmission.In order to protect the security of images during transmission,image encryption technology Research has become a top priority.The chaotic system has the characteristic of being very sensitive to the initial value,which is very beneficial to the research of image encryption,so it is one of the most commonly used methods to study image encryption,but the encryption technology is not perfect.In order to improve the simplification of the existing encryption technology,the lack of key space and the degradation of chaotic dynamics,this paper carries out the following research in conjunction with the chaotic system.(1)A new replacement method is proposed for image encryption.First,the image is replaced by the pixel value position of the image through cat mapping,and secondly,the new matrix magic transformation(MMT)replacement method is used to completely disrupt the image pixel value.The MMT replacement is to generate an index matrix through a chaotic sequence matrix,and then perform a replacement operation on the image according to the coordinate position indicated by the index matrix.The perfect combination of cat mapping and MMT replacement method not only eliminates the periodic shortcomings that occur when cat mapping is used alone,but also greatly improves the confusion of image pixel values.In the diffusion process,the adjacent xor operation of the permutation matrix and the chaotic sequence matrix is adopted,and the key diffusion is added to enhance the encryption degree of the original image.(2)A bit-level image encryption algorithm based on Back Propagation(BP)neural network and Gray code is proposed.First,the plaintext image is decomposed into binary images,and then the Lorenz hyperchaotic system is used to generate two sets of chaotic sequences,which are used for Gray code bit-level permutation operations to generate permutation matrices.During diffusion,the permutation matrix is converted into a bit matrix and output in reverse order to generate a diffusion matrix.The key stream is generated by the BP neural network.The key stream and the diffusion matrix are subjected to the neighbor exclusive xor operation,which increases the robustness of the ciphertext.(3)An image encryption algorithm based on spatiotemporal chaos and binary tree in-order traversal is proposed.The main simulation image is usually 256×256,so the midorder traversal is performed on the values of 1-65536,as an index matrix,the plaintext image is arranged globally according to the mid-order traversal sequence on the permutation,and then the spatiotemporal chaotic system is used to generate the chaotic sequence,and the chaotic interference is set value,complete the xor operation between the adjacent bits of the replacement image,generate the ciphertext matrix,and complete the image encryption process.Experimental analysis shows that this algorithm has a good encryption effect.The three encryption algorithms generate chaotic sequences based on different chaotic systems,and all make innovations in permutation.Experimental analysis shows that the three permutation methods have better encryption effects than traditional permutation methods.