COMPARATIVE ANALYSIS OF IMAGE STEGANOGRAPHY TECHNIQUES BASED ON NEURAL NETWORK AND SVM

Jagpreet Kaur

Abstract


Steganography is an ancient technique that involves hiding secret data in images for covert communications. This technique aims at hiding the secret message in an image in such a way that the very presence of message is not obvious. In this paper, a comparative analysis of image steganography techniques based on various machine learning techniques have been done. Image segmentation using Discrete Wavelet Transform forms the basis of this technique. Image segmentation has been incorporated to provide the basic segments in the cover image where data can be embedded. The optimal bit positions for information embedding are decided by Neural Network and Support Vector Machine based on their training with a set of images. Firstly DWT will be implemented to do segmentation. Then training of images will be done using Neural Network and Support Vector Machine. After that message embedding will be done using SVM and as well as NN and message extraction  process  also involves use of both SVM and NN. The quantitative evaluation have been done using MSE, PSNR and time complexity in MATLAB 7.10 environment. It has been empirically concluded that NN based image stegangraphy is more secure and faster than SVM based steganography.

KEYWORDS: Steganography, DWT, Neural Network, SVM.


Full Text:

PDF2

Refbacks

  • There are currently no refbacks.


© 2017 International Journal of Global Research in Computer Science (JGRCS)
Copyright Agreement & Authorship Responsibility