Color Image Encryption Algorithm Based on Hopfield Chaotic Neural Network
Liping Liu · International Journal of Computational Intelligence and Applications · 2026
In view of the shortcomings of current image encryption algorithms in key space, encryption efficiency and anti-attack robustness, a multi-component color image encryption algorithm (L-ACM-HNN-CS-DNA algorithm for short) based on Hopfield chaotic neural network and combined with Logistic mapping, is proposed. Logistic mapping provides initial randomness and drives pixel scrambling. Hopfield network generates a diffusion matrix to achieve pixel value confusion. Compressed sensing technology is introduced to optimize data compression and transmission efficiency, and DNA encoding technology is used to enhance diffusion security. Experimental results showed that in terms of encryption efficiency, when the image size was [Formula: see text], the encryption throughput reached 1170[Formula: see text]MB/s. In terms of key sensitivity, when the control parameter was [Formula: see text], the pixel change rate reached 99.72%. In actual application, in terms of statistical characteristics, the encrypted image information entropy reached 7.995. In terms of attack resistance, the average change intensity index was stable at 33.65%. In terms of robustness, the anti-cropping ability still maintained 0.52 under a 35% cropping ratio. The results show that the algorithm performs well in security, efficiency and robustness, and provides a new security solution for color image encryption.