1. Python 3.6.x
2. OpenCV 3.x

Note: Make sure that Python is compiled with framework enabled when installing on macOS systems.

This project requires a working installation of OpenCV 3. Please install this first before installing the project dependencies.

Dependencies of this project can be installed via PIP.

Follow these steps to install the setup the application:

1. Run the command pyvenv venv to setup a Python 3 virtual environment.

2. Activate the virtual environment by running source venv/bin/activate

3. Install the project dependencies by running pip install -r requirements.txt

4. Locate your OpenCV Python bindings and type the command echo <Python OpenCV bindings path> >> ./venv/lib/<python version>/site-packages/opencv3.pth.

   Where:

   1. <Python CV bindings path>: is the site-packages folder inside the OpenCV installation. Please make sure you select the appropriate version of bindings that matches the Python version declared in the requirements.

   2. <python version>: is the version under your venv virtual environment folder

5. Create the folder assets/img inside the root of the project directory and place the input images inside the newly created folder.

You can train the OCR by running the command python train.py or using the command line tool provided by typing the command ./ocr.py train <training_images_directory> <output_knowledge_data_path>. For more options in the command line tool, you can use ./ocr.py train --help.

You can verify the ability of the OCR to recognize the characters from the test image by running python test.py or using the command line tool provided by typing the command ./ocr.py recognize <input_image> <knowledge_data_path>. For more options in the command line tool, you can use ./ocr.py recognize --help.

MIT