samhooke / Steganography

Image & video steganography in Matlab
42 stars 36 forks source link
steganography

About

This code was written for a project titled "Applying Steganography to Digital Media".

Setup

In Matlab, do: File -> Set Path... -> Add With Subfolders... -> Browse for root directory of this repo -> OK

Steganography

This project contains implementations of 6 distinct steganography algorithms in Matlab, a simple interface for applying steganography to images, and a couple of processing methods for increasing robustness of the algorithms. For each algorithm, there is an example application both images and videos. The following files are all located in the /steganography folder:

Algorithm Image Example Video Example
LSB steganography_image_lsb.m steganography_video.m
DCT steganography_image_dct.m steganography_video.m
ZK steganography_image_zk.m steganography_video.m
WDCT steganography_image_wdct.m steganography_video.m
Fusion steganography_image_fusion.m steganography_video.m
Egypt steganography_image_egypt.m steganography_video.m

After running any of the above files, statistics about the algorithm performance are displayed. They are generated by steganography_statistics.m, which contains many customisable options.

A basic spellchecker, written in Python, is also included. It can optionally be applied through modifying steganography_statistics.m.

Hamming codes can also be applied. They can be enabled in the DCT or Egypt algorithms through modifiying their use_hamming parameter.

Interface

Within the /interface folder exists ImageInterface.m. This is a simple Matlab interface that allows performing steganography of text into images using all of the above algorithms.

It is intended only for demonstration purposes. Much more customisation of the algorithms is possible through performing steganography directly through Matlab code.

Input, Output and Results

The /input folder contains potential carrier images, such as the infamous lena.jpg.

The /output folder is where the steganography algorithms output their stego-images when they are run for 1 iteration.

The /results folder is where the steganography algorithms output their stego-images and a CSV of results when they are run for multiple iterations.

The number of iterations is set by modifying the iteration_total parameter at the top of each steganography example file.

Other Parameters

Each algorithm can be switched between greyscale and colour, through changing use_greyscale. If using colour, the specific colour channel can be set with channel.

Code

This project is about 98% written in Matlab, with 2% Python. If Python is not installed the Matlab code will still run succesfully, but the Spellchecker will not function, and message_similarity_py will return NaN.