Directory OCR

This project extracts structured data from historical directory images using OCR and Python. This is not a complete project, just a showcase

Requirements

• Python 3.8+
• Tesseract OCR (installed and in your PATH)
• The following Python packages:
  • opencv-python
  • numpy
  • pytesseract
  • Pillow
  • pandas
  • spacy

Setup

1. Clone the repository and navigate to the project directory.

2. Install Python dependencies:

   pip install -r requirements.txt

3. Install the spaCy English model:

   python -m spacy download en_core_web_sm

4. Install Tesseract OCR:

   • Windows: Download here and add the install directory (e.g., C:\Program Files\Tesseract-OCR) to your PATH.
   • macOS:

     brew install tesseract

   • Linux:

     sudo apt-get install tesseract-ocr

Usage

1. Place your directory images in the sample-data folder.

   • Update the image path in directory_ocr.py if needed.

2. Run the script:

   python directory_ocr.py

3. Output:

   • The script prints parsed entries to the console.
   • It saves a CSV file named test_output.csv with columns: last, first, occupation, home_addr.