RGB image Encryption and Decryption with Neural Chaotic Functions

It is necessary to protect images from different people due to the fact that images have very sensitive and confidential information in different applications. Encryption is one of the best ways to store information. Encryption involves the process of converting information using an algorithm, so that the encrypted information is unreadable to all, except authorized users with special knowledge who actually have the keywords. This research presents a fast and efficient method with the best rate of peak signal-to-noise ratio compared to previous methods, so that it offers a combination of deep neural network based on chaotic functions. Training and testing in encryption is a difficult which considered in recent years due to sensitive mode, therefore, the Neocognitron deep neural network is considered and for the placement and permutation of image pixels for the purpose of encryption, the chaos function named Fractal is used. Also, sending a key to read the image is essential which is called a decryption operation. This research deals with cryptographic operations with decoding by combining Neocognitron deep neural network and fractal model. Due to the fact that chaotic systems have complex and ergodic behavior and also have a quasi-random structure that is very sensitive to the initial values and conditions, so they can be used in image encryption. The results of the proposed approach indicate that the evaluation criteria have improved in compare of the previous similar methods.

کلیدواژه ها:

Image Encryption ، Image Decryption ، Deep Neural Network ، Neocognitron ، chaotic Function ، Fractal