A Novel Image Encryption Algorithm Based on Cellular Neural Networks Hyper Chaotic System
Gangyi Hu, Weili Kou, Jian-E Dong, Jin Peng · 2018
Image encryption based on chaos has been extensively researched in the past, many researchers have approached the problem of dynamic key from the image pixels, however, such key can be easily shuffled when the cipher image has been polluted or damaged through the transmission, in which the decryption becomes very difficult. Some recent researches have been trying to solve this problem, but they failed to achieve high security. This paper proposes an image encryption algorithm based on CNN (Cellular Neural Networks) hyper chaotic system, and the image encryption key is encrypted by using the asymmetric RSA algorithm. Our algorithm uses the key as input through the CNN system to generate six dimensional chaotic sequences. Then we change the original image pixel positions by pair-wise sorting the pixels with the values of the chaotic sequences. Finally replace the pixel values by performing XOR operation between the pixel values and the sequences to get the cipher image. The simulation experiments show that, compared with other chaotic image encryption schemes, it can resist the transmission noise. Moreover, it can also decrypt the image even after some part of the cipher image had been lost. The cipher image can also achieve high change rate of pixel ratio, high information entropy, and strong anti-hacking ability.