python3 main.py [<options>] <image_path> [<image_path> [...]]
python3 main.py --help
Options:
-s, --silent: do not show any windows, print output only
--json: output in json format
-h, --help: show this help message
Install python3
pip install opencv-contrib-python # 4.1.0.25
pip install imutils # 0.5.2
pip install Pillow # 3.1.2
pip install pytesseract # 0.2.6# Install tesseract version 4
# https://launchpad.net/~alex-p/+archive/ubuntu/tesseract-ocr
sudo add-apt-repository ppa:alex-p/tesseract-ocr
sudo apt-get update
sudo apt install tesseract-ocr # 4.1.0~git3875-b2fc3eba-1ppa1~xenial1- Go to https://github.com/Shreeshrii/tessdata_arabic
- Download these files to your system's
tessdatadirectory (in Ubunutu:/usr/share/tesseract-ocr/4.00/tessdata)- ara-Amiri.traineddata
- ara-Amiri-layer.traineddata
- ara-Scheherazade.traineddata
Note: This project uses tesseract for OCR, with custom trained data for arabic taken from here, because default trained data does not recognize the hindi numerals (see this issue comment for more details)