Design Of Image Encryption System Based On Chaotic Mapping And Hopfield Neural Network And FPGA Implementatio

With the evolution of communication networks and streaming media technologies,information leakage occurred more frequently.Information security has become an unignorable public issue.Image security has reflected in all aspects of society.The need to guarantee information security has prompted image security technologies to be one of the focal points.Existed encryption schemes may have problems such as an insufficient key space,non-one-time pad,or easily degraded simple structure.To solve these problems,this paper has proposed an image encryption algorithm based on chaotic map and Hopfield neural network,designed and implemented an image encryption system on FPGA platform as well.In order to obtain a large enough key space,this paper united Logistic map and highdimensional Hopfield neural network to construct a pseudo-random sequence generator,generating key sequences by iteration to achieve the requirement of the key space security.For the one-time-pad encryption problem,this paper designed a method to generate the initial key according to the image chunking and use the selected key to index in the key table,to get the initial keys of multiple systems respectively,which has further expanded the key space of the algorithm on the foundation of plaintext correlation.For the problem of encryption structure,a permutation-dynamic diffusion encryption method was designed in this paper,which has enhanced the complexity of the encryption structure.Compared with other referenced methods in the paper,the proposed encryption algorithm possessed higher security metrics and had the ability to resist typical attacks.In the digital circuit design of the image encryption system,this paper used an iterative CORDIC structure replacing hyperbolic function and divider to implement a neural network activation function with enable end,achieving high accuracy.This paper completed the FPGA architecture design of the high-dimensional Hopfield neural network,laying the foundation for the application of the network in the designed neural network encryption FPGA chip.Based on the Hopfield neural network and Logistic map pseudo-random sequence generator,the Arnold permutation algorithm was implemented by coordinate transformation,and the dynamic diffusion encryption algorithm was further completed by quantization of the key sequences.Finally,the image encryption and decryption results were demonstrated on the LCD display.The analysis of the resource consumption and security of the system has shown that the designed system didn’t consume too many resources and was able to meet the security requirements.