Document Intelligence with OpenCV & Tesseract OCR

A cross-platform guide for document text extraction using OpenCV and Tesseract OCR. Complete with OS-specific setup instructions.

---

🙌 Intro

A beginner-friendly guide to extracting text from documents using OpenCV for image processing and Tesseract OCR for text recognition. Designed for AI/Engineering students to understand foundational document intelligence concepts.

---

📖 Purpose

This project demonstrates how to:

• Detect text regions in images using basic computer vision techniques
• Extract machine-readable text with Tesseract OCR
• Build interactive document scanners with Streamlit/Jupyter
• No deep learning or complex libraries (like LayoutParser) required!

---

🔑 Key Takeaways

1. OCR Workflow: Preprocessing → Text Localization → OCR → Postprocessing
2. Tool Roles:
   • OpenCV: Image thresholding, contour detection, ROI extraction
   • Tesseract: Optical Character Recognition (OCR) engine
3. Real-World Challenges: Handling low contrast, complex layouts, multi-language text
4. Limitations: Simpler but less accurate than deep learning approaches (e.g., LayoutParser)

---

🖥️ System Requirements

All Platforms

• Python 3.8+
• 4GB RAM (minimum)
• 500MB disk space

macOS Specific

• macOS 10.15 (Catalina) or newer
• Xcode Command Line Tools

Linux Specific

• Ubuntu 20.04/Debian 10 or equivalent
• GTK+ 3.x libraries

---

🛠️ Platform-Specific Setup

1. Tesseract OCR Installation

OS	Command	Additional Notes
Windows	Download installer	Check "Add to PATH" during install
macOS	brew install tesseract	Requires Homebrew
Linux	sudo apt install tesseract-ocr libtesseract-dev	For Debian/Ubuntu

2. System Dependencies

macOS:

# Install Homebrew if missing
/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"

# Install image libs
brew install leptonica

---

⚙️ Configuration Guide

Tesseract Path Setup

Add this to your Python code:

import pytesseract

# Windows
pytesseract.pytesseract.tesseract_cmd = r'C:\Program Files\Tesseract-OCR\tesseract.exe'

# macOS (Homebrew install)
pytesseract.pytesseract.tesseract_cmd = '/usr/local/bin/tesseract'

# Linux
pytesseract.pytesseract.tesseract_cmd = '/usr/bin/tesseract'

---

🐧 Linux-Specific Notes

1. Window Manager Conflicts:
   If using headless Linux server:

   sudo apt install xvfb
   export DISPLAY=:0

2. Font Issues: Install additional fonts

   sudo apt install tesseract-ocr-eng tesseract-ocr-fra  # etc.

---

 macOS-Specific Notes

1. M1/M2 Chip Optimization:
   Use native ARM Homebrew in Terminal:

   arch -arm64 brew install tesseract

2. Gatekeeper Issues: If blocked by macOS security:

   xattr -d com.apple.quarantine /path/to/tesseract

---

🚀 Universal Installation

# Clone repo
git clone https://github.com/yourusername/document-intelligence-demo.git
cd document-intelligence-demo

# Install requirements (in virtual env)
pip install -r requirements.txt

---

▶️ Running the Project

All OS:

streamlit run app.py  # Web app
jupyter notebook      # Jupyter version

---

🚨 Troubleshooting

OS	Issue	Solution
macOS	Error: Failed building wheel for opencv	brew install cmake pkg-config
Linux	ImportError: libGL.so.1	sudo apt install libgl1-mesa-glx
All	TesseractNotFoundError	Verify path with which tesseract

---

📜 License

MIT License - Free for academic and commercial use. Tesseract OCR is Apache 2.0 licensed.