Secure Image Steganography Algorithm Using Radial Basis Function Neural Network

Abstract

Recently, ensuring the security of secret messages over computer networks has significantly increased in importance. For this reason, a new system is proposed that tries to hide text using Artificial Neural Network (ANN), and more precisely using Radial Based Function, with zero mean square error, in addition to encryption techniques, to make sure that the resulting text is exactly the same as the one that was sent. In this study the text is encrypted by an ordinary encryption algorithm, then the encrypted text will be embedded within the image and the positions of each encrypted text value will be determined, and in the last step the taken values (positions) will be encrypted using the neural network. The resulting encrypted text is unpredictable, making it very secure. On the receiver side, only the person, who has knowledge of the decryption key, neural network inputs P and parameters, will be able to see the original message embedded in the image.