The Memorable Image Encryption Algorithm Based on Neuron-Like Scheme

A new image cryptosystem based on neuron-like scheme is proposed in this paper. As a non-linear component of image cryptosystem, neuron-like learning algorithm can realize a kind of memorable diffusion algorithm, thus realizing the plaintext-related image encryption algorithm. Through the feedback operation, the inputs and weights of the neuron are regulated by the information of the image, and then they are used to diffuse the image information. According to the needs of application scenarios, the number of inputs of the neuron can be adjusted so that the image cryptosystem can be applied to different security requirements. The simulation results show that the proposed image encryption system is extremely sensitive to the secret keys, the equivalent keys and the plain images, and also possesses the high security and adaptive characteristics. The image cryptosystem can be applied to the actual applications of image information security.