Cyber Times International Journal of Technology & Management

Neural Cryptography, the latest emerging trend of cryptocysm whirls around an approach of eliminating the need of a secure public channel by generating a key on the sender’s as well as the receiver’s side so as to enhance the overall security of a cryptographic system. Image Steganography being a derived rendition of cryptography embeds messages in an image and shares the key over a channel for the decryption while using a symmetric key algorithm thus making it easy for an adversary eavesdropper to get the key and decrypt the message. In this paper a novel approach of symmetric key encryption based on AES-256 bit PRP and Pseudo Random Hash based Neural Cryptographic LSB technique has been proposed to select the bits’ insertion position of secret information for image steganography. In order to increase the security of the communication system, we make the use of Multi Layered Perceptron Neural Networks by using the back propagation algorithm for generating the key and thus eliminating the need for the communication channel. The proposed method is analyzed in terms of Image Fidelity and results are found to be promoting.