Research On Digital Image Encryption Based On Chaos And Neural Network

With the development of computer technology,the information security through the Internet has become more and more concerned by people.As one of the main forms of multimedia communication,the security of digital images has become the focus and difficulty in information protection.This paper will focus on the application of chaos and neural network in digital image encryption.It theoretically analyzes and innovates chaotic theory and neural network,and successfully applies it to the field of image encryption.Its encryption effect is good and its security is high.In the current research,with the rise of artificial intelligence,the research of neural networks has entered a period of rapid development.The neural network has excellent performance of associative memory and processing optimization.Such traits are in line with the demand for the diffusion of plaintext information in chaotic cryptography.However,in the process of combining neural network and chaotic cryptography,most of the chaotic systems used are relatively simple or have no application process,which may cause certain security risks.Based on these issues,this paper has carried out the following work:(1)A chaotic image encryption algorithm combining stream ciphers and block ciphers is proposed.That is,the stream cipher is combined with the block cipher to spread,and then the logistic mapping matrix is used for the block sorting transformation to complete the scrambling.Combining stream encryption with packet encryption,using S-box and shift operation,for the replacement and processing of keys,the diffusion of plaintext information is completed,and then the semi-ciphertext is scrambled by block sorting to get the final result.Cipher text.The final security analysis can show that it has good security and can resist common attacks.(2)A chaotic color image encryption algorithm based on Hopfield chaotic neural network is proposed.Firstly,a new phased composite chaotic map combining staged Logistic map and Tent map is proposed and used for scrambling.The comparative study shows that the map cryptographic characteristics are better than Logistic map.Secondly,Hopfield chaos is used.The neural network generates a self-diffusion chaotic matrix for the diffusion of semi-ciphertext to complete the final encryption.Compared with other experimental results,it can be seen that the encryption efficiency is improved and the security is guaranteed.(3)A new image encryption algorithm based on fractional-order 5-dimensional cellular neural network and Fisher-Yates scrambling is proposed.Firstly,the research of fractional-order 5-dimensional cellular neural network system is proposed and carried out.With its higher complexity,it is used as the diffusion controller in the encryption process.Secondly,the complex chaotic map is used as the Fisher-Yates scrambling controller to scramble the ciphertext and then we can get the final ciphertext.