readme.md

Preparing the publicly available datasets

Below we detail the instructions to use the two publicly available browser fingerprint datasets.

FPStalker

The FPStalker dataset is available on the FPStalker github repository.

1. Download the two archives extension1.txt.tar.gz and extension2.txt.tar.gz.
2. Extract the two archives and regroup them in a file named tableFingerprints.sql:

tar zxvf extension1.txt.tar.gz
tar zxvf extension2.txt.tar.gz
cat extension1.txt extension2.txt > tableFingerprints.sql

3. Prepare the FPStalker dataset using the dedicated preprocessing script:

python -m executables.dataset.preprocess_fpstalker sql_script_path output_directory

• The sql_script_path is the path to the tableFingerprints.sql file (which is required to have this name).
• The output_directory is the directory where to save the resulting data.
• You can specify the --keep-raw-canvas option to hold the complete base64 encoded canvas images in the fingerprint dataset.

4. The resulting browser fingerprint dataset is in output_directory:

• The fingerprint dataset is saved in fingerprints.csv.
• The canvases are extracted as images in the canvas directory with a subdirectory named after the type of canvas and each image is named after its hash value in the fingerprint dataset.

HTillmann

The HTillmann dataset is available on the blog of Henning Tillmann.

1. Download the file named MySQL-Tabelle “bfp”.
2. Extract the archive as a file named bfp.sql:

unzip bfp.sql.zip

3. Prepare the HTillmann dataset using the dedicated preprocessing script:

python -m executables.dataset.preprocess_htillmann sql_script_path output_directory

• The sql_script_path is the path to the bfp.sql file (which is required to have this name).
• The output_directory is the directory where to save the resulting data.

4. The resulting browser fingerprint dataset is saved in output_directory/fingerprints.csv.