Information Hiding in RGB Images Using an Improved Matrix Pattern Approach

Amirfarhad Nilizadeh; Wojciech Mazurczyk; Cliff Zou; Gary T. Leavens; Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, 2017, pp. 103-111

Abstract


In this paper, a novel steganography algorithm based on an improved "Matrix Pattern" (MP) method is presented. In this process, firstly, an RGB image is divided into the non-overlapping square-sized blocks. Next, 95 dynamic-sized unique matrix patterns are automatically generated using the 4th and 5th bit layers of the green layer of each block, which are assigned to 95 English keyboard characters. Then, the blue layer of each block is used for embedding secret messages by adding matrix patterns which are assigned to the characters of the secret message. The results show that this algorithm has a high resistance against steganalysis attacks, including Regular Singular (RS), Sample Pair (SP), and PVD based attacks. Furthermore, the proposed algorithm not only improves capacity by over 27% when compared to the existing method, but also results in a slightly better transparency of the stego-image.

Related Material


[pdf]
[bibtex]
@InProceedings{Leavens_2017_CVPR_Workshops,
author = {Nilizadeh; Wojciech Mazurczyk; Cliff Zou; Gary Leavens, Amirfarhad T.},
title = {Information Hiding in RGB Images Using an Improved Matrix Pattern Approach},
booktitle = {Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshops},
month = {July},
year = {2017}
}