ING Contest: who writes the best document reading tool
In this contest you are gonna play with opensource OCR tooling. The contest is based on real project. Due to migration of old archives to a new archive, documents where stored under wrong contractnumbers. Before putting the documents in the target archive the correct documentsnumbers must be extracted from the contract using OCR tooling.
Unfortunately all documents were not scanned very accurate, so you need to do some tricks to retrieve the right contractnumbers. Can you write a DocumentProcessor class that recognises all the right contractnumbers?
You can test your DocumentProcessor using the testSuite.
Please feel free not to use Tesseract but use your own tools.
Good luck!
To run this programm you need to first: Windows:
- install the Tesseract software (https://youtu.be/QJkKDsjj1oA) - windows
- create a virtual environment
python3 -m venv env - activate the virtual environment
env\Scripts\activate.bat - install dependencies
pip install -r Requirements.txt - make sure in the DocumentProcessor the path to the Tesseract executable is correct
Mac:
- Make sure python3 is installed (this is the esiest using
brew install python) - install the Tesseract software using homebrew:
brew install tesseract- mac users - Create a virtual environment
python3 -m venv env - Activate the virtual environment
source env/bin/activate - install dependencies
pip install -r Requirements.txt - If you don't have brew installed in the default prefix, add the following environment variable:
export TESSDATA_PREFIX=$(brew --prefix)/share/tessdata
Linux Uitgegaan van Raspbian/VSCode (Debian)
- sudo apt install tesseract-ocr y
- sudo apt install tesseract-ocr-dut … make a programming directory … go there
- git clone https://github.com/gertvermeer/ingcontest.git … in VSCode
- install Pylance etc.