AMR-KELEG / Digitalizing-Arabic-Documents

Arabic OCR using Tesseract

This repository serves as a guide for digitalizing Arabic documents using Tesseract.

Steps

• Install the python packages using pip pip install -r requirements.txt

• Convert the pdf file into multiple pngs mkdir -p pngs && convert -density 150 -trim PDF_FILE.pdf pngs/page%d.png.

  • Note: Replace PDF_FILE.pdf with your file's name.

• Download the Arabic model to the tessdata/ directory from the tessdata repository: https://github.com/tesseract-ocr/tessdata/blob/main/ara.traineddata

• Run the OCR generation script: python run_ocr.py.

• Inspect the output tsv file and the input pdf file for ad-hoc text preprocessing operations to clean the text.

Archive of Datasets Generated Using the Pipeline

• Parallel Tunisian Constitution Corpus (PTCC):