Neuronal crypto system based on chaotic super-increasing sequence

Neural networks were originally developed as an alternative to artificial intelligence to solve problems such as control, shape or word recognition, decision-making, memory, and more. In this work, we propose to assign them to secure images. For this, we use data (which is defined from the logistics map) to generate a super-increasing sequence which can attributed as a weight of the synapses. Also, we have inspired to encrypt the images by the Merkel-Hellman algorithm. The proposed system is called Neuronal Crypto-System which is a secret key system. It takes some steps of encryption and decryption of a specific public key crypto system and crypts a grayscale image to color image (RGB).