Image and Audio Embedding Technique in Image Steganography Using Neural Networks

Abstract: In this paper, we clarify what steganography is and what it can do. The Steganography is the prowess of hiding information in the ways that fend off the revealing of secret messages. This project work discusses about image and audio embedding technique based on Neural Network FFBP-NN. Feed forward back propagation technique enhances the surety of the data. The data is encrypted by using the DCT technique and then hidden using the medical image and audio by tapping the abilities of FFBP-NN with the use of linear embedding technique and considering the characteristics of Human Audio System (HAS), a NN restrainer(controller) is designed to ensure the strength of embedded data adapting to the host audio itself entirely. The simulation experiment results show that the technique is racy (robust) too common digital audio processing methods as well as the quality of the medical image and audio is guaranteed. This paper mainly focused on the problem of audio and the image using Artificial Neural Networks (ANNs) has been addressed. Neural Network is trained to recognize/classify elementary-actions such as epochs, time, gradient, performance and quality of the image. The applications being used in this project mainly concentrate on FFBP-NN. Steganography system using features derived from Discrete Cosine Transforms (DCT) coefficients along with FFBP-NN classifier is evaluated, using an image dataset of thirty images, containing four classes and each class having five images. The action being performed is classified and displayed on the user interface along with a spoken sound version with the help of “Levenberg-Marquardtm” method.

Keywords: DCT, IDCT, Linear Embedding, Encryption Technique, FFBP-NN, Blocks, Co-efficient, RGB image.