About

Document Forger is a Python package that allows you to create a custom-defined number of documents using one document. This package generates as many forged or synthetic documents as the user needs.

Installation

Use the package manager pip to install document forger.

python setup.py install

Or clone this git repository and install the dependencies through the requirements text file provided.

pip install -r requirements.txt

How it Works

The package is built around using the copy-paste technique. The code utilizes OCR to detect and recognize words and their bounding boxes. The code then goes through the words and decides whether or not two characters are swappable. If they are, the code swaps the first character with the second character. This allows us to create minor forgeries that are unrecognizable to the naked eye but still obvious enough to detection software and AIs. The purpose behind this package is to artificially expand and create a synthetic dataset that can be used to test Forgery Detection AI and to stress test it with different variations.

Alternate Text
Alternate Text

The above images shows real vs forged generated documents where the red boxes highlight the modifications made to the real document. This shows the capabilities of our package to work with different sizes, styles and fonts.

Usage

Through Scripts:

from document_forger.document_processing import process_document

process_document(input_image, output_directory)

Or through the terminal

python -m document_forger --image_path input_img --ouptut_dir output_dir

If you have tesseract installed and added to your local enviorments, than set the path to the exe using the following:

from document_forger.utils import set_tesseract_cmd

set_tesseract_cmd(exe_path)

Or you can use the --tesseract_cmd  argument in the terminal.

View More research

vfusedSeg3d

An innovative multimodal fusion system created that combines camera and LiDAR data to significantly enhance the accuracy of 3D perception.

Learn More

Forgery Detection

An advance AI approach to detect forged documents.

Learn More

VideoGrad Vis

An advanced package that encompasses cutting-edge methods for Explainable AI (XAI) in the realm of computer vision.

Learn More

Take the next Step

Contact Sales

For product demonstration or collaboration, please contact us here.

Request Information

For all other questions, please contact us here.