This project is created for my bachelor thesis "Creating a Detection Tool for Data Hiding Techniques
in EXT4" at University Leiden, Netherlands. The bachelor thesis was supervised by dr. K. F. D. Rietveld.
This project contains a detection, hide and benchmark tool, all written in Python. All the tools require Python 3.10. All the tools are CLI programs,
which means they can be used from the command line.
To hide data, the following command can be used:
python3 Hide.py -f path/to/image.dd -d dataToHide -t techniqueNrToUse --log/--no-log
All the techniques used to hide data are listed in the thesis, which can be found ... .
To detect any hidden data, the following command can be used:
python3 Detect.py -f path/to/image.dd (-s string) --log/-no-log
With the -s command, one can explicity search for a string, and the program will only report data found which
has the requested string in it.
To run the benchmark, the following command can be used:
python3 BenchmarkDHEXT4.py -i path/to/imagecatalog.xml -t path/to/techniquecatalog.xml --search/--no-search
The image and technique catalog can be generated using the generator program found in the Catalog map. With --search/--no-search,
one can run the benchmark while the detection tool searches for a string, or the detection tool searches for the 'unknown'.
It is possible to create your own Hide and Detection class to work with the benchmark. These classes have a few required methods and initializers to work with the benchmark tool:
The detection class requires the following initalizor: Detect(file_name, log, string). Hereby is file_name a String which obtains the file name of the image, log is a Boolean which controls the logging output and string is a list of Strings, which contains words to search for. If string is None, the program searches for the 'unknown'
Furthermore, a method called check_all() is required, which calls all the detection methods implemented in the class. This method returns a list of Strings, with the names of all found data hiding techniques.
The hiding class requires the following initalizor: Hide.Hide(file_name, type, data, inode, group). Hereby is file_name a String which obtains the file name of the image, type the hiding technique, data the data which needs to be hidden, inode the inode number of where the data must be hidden and group the group number of where the data must be hidden.
Furthermore, a method called check_if_possible() is required, which checks if it is possible to perform the given data hiding technique on the image (f.e., check if block size is > 1024 for Superblock Slack).
At last, a method called get_hiding_technique() is required. This method calls the right hiding method based on the 'Type' input, and performs the hiding technique on the image.
By changing the Detect import in BenchmarkDHEXT4.py and the Hide import in Catalog/parser.py, one can now use it's own implementation of a Detection and Hide class for the benchmark