https://github.com/shahin-ro/table-detection
Python tool for table extraction & Persian OCR. Uses OpenCV for table detection, Tesseract for text extraction, & Pandas for data output. Visualizes cells & text. Ideal for Persian documents! 📄✨
https://github.com/shahin-ro/table-detection
colab computer-vision data-extraction data-visualization document-processing image-analysis image-processing machine-learning matplotlib numpy ocr opencv pandas persian-ocr persian-text python table-detection table-extraction tesseract text-recognition
Last synced: 2 days ago
JSON representation
Python tool for table extraction & Persian OCR. Uses OpenCV for table detection, Tesseract for text extraction, & Pandas for data output. Visualizes cells & text. Ideal for Persian documents! 📄✨
- Host: GitHub
- URL: https://github.com/shahin-ro/table-detection
- Owner: shahin-ro
- License: mit
- Created: 2025-08-17T06:38:12.000Z (8 months ago)
- Default Branch: main
- Last Pushed: 2025-08-17T07:13:05.000Z (8 months ago)
- Last Synced: 2025-08-17T08:33:26.601Z (8 months ago)
- Topics: colab, computer-vision, data-extraction, data-visualization, document-processing, image-analysis, image-processing, machine-learning, matplotlib, numpy, ocr, opencv, pandas, persian-ocr, persian-text, python, table-detection, table-extraction, tesseract, text-recognition
- Language: Jupyter Notebook
- Homepage:
- Size: 266 KB
- Stars: 1
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Table-Detection
Table Extraction and OCR for Persian Documents 📄✨
This project provides a Python-based solution for detecting table
structures in images and extracting Persian text using Optical Character
Recognition (OCR). It uses OpenCV for table detection and Tesseract OCR
for text extraction, with proper rendering of Persian text. 🚀
## Features 🌟
- **Table Detection** 📊: Identifies table cells in images using
advanced image processing with OpenCV.
- **OCR Support** 🔍: Extracts Persian text from table cells using
Tesseract OCR with Persian language support.
- **Data Structuring** 📈: Organizes extracted text into a Pandas
DataFrame for easy analysis.
- **Visualization** 🎨: Displays detected table cells, line masks, and
intersection points for verification.
## Requirements 🛠️
To run this project, you need the following dependencies: - Python 3.7+
🐍 - OpenCV (`cv2`) - NumPy - Matplotlib - Pandas - Pytesseract (for
Tesseract OCR) - Tesseract-OCR with Persian language support
(`tesseract-ocr-fas`)
Install the dependencies using:
``` bash
pip install opencv-python numpy matplotlib pandas pytesseract
```
For Tesseract OCR:
``` bash
apt-get install -y tesseract-ocr tesseract-ocr-fas
```
## Usage 🚀
1. **Clone the Repository** 📂:
``` bash
git clone https://github.com/shahin-ro/table-extraction-ocr.git
cd table-extraction-ocr
```
2. **Prepare an Image** 🖼️:
- Ensure you have an image containing a table with Persian text
(e.g., a scanned document or screenshot).
- Place the image in the project directory or provide the path to
the script.
3. **Run the Script** ▶️:
- The script (`jadval.py`) processes the image, detects table
cells, extracts text, and visualizes the results.
- Run the script:
``` bash
python jadval.py
```
4. **Output** 📜:
- The script outputs:
- A count of detected table cells ✅.
- Extracted text for each cell with coordinates 📍.
- A Pandas DataFrame representing the table structure 🗃️.
- Visualizations showing detected cells, line masks, and
intersection points 🖼️.
## How It Works 🧠
1. **Table Detection** 📏:
- Uses OpenCV to preprocess the image (grayscale, adaptive
thresholding, morphological operations).
- Detects horizontal and vertical lines to identify table
boundaries.
- Clusters line intersections to determine cell coordinates.
2. **Text Extraction** 📝:
- Crops each detected cell and processes it with Tesseract OCR
(`lang='fas'`) for Persian text extraction.
- Stores text and coordinates for each cell.
3. **Data Structuring** 📚:
- Maps extracted text to a grid based on cell positions.
- Creates a Pandas DataFrame to represent the table structure.
4. **Visualization** 🖌️:
- Displays three plots:
- **Detected Cells** 🟢: Original image with green rectangles
around table cells.
- **Line Mask** ⚪: Inverted mask showing detected horizontal
and vertical lines.
- **Joints** 🔲: Intersection points of table lines.
## Example 📋
``` python
# Example output for a table with 6 cells
✅ Detected 6 cells.
متن سلول 1: نام
مختصات: (50, 30, 150, 80)
---
متن سلول 2: سن
مختصات: (150, 30, 250, 80)
---
...
جدول استخراج شده (متن داخل سلولها):
0 1 2
0 نام سن شغل
1 علی 30 مهندس
```
## Notes 📌
- **Tesseract OCR** 🔍: Requires `tesseract-ocr-fas` for Persian
language support.
- **Colab Compatibility** ☁️: The script is designed to work in Google
Colab, with file upload support and Tesseract installation commands.
- **Image Quality** 🖼️: OCR accuracy depends on clear table lines and
readable text.
## Limitations ⚠️
- The table detection algorithm assumes well-defined table lines.
- OCR accuracy depends on image quality and text clarity.
- Persian text rendering in visualizations may require additional font
support for non-Colab environments.
## Contributing 🤝
Contributions are welcome! Please submit a pull request or open an issue
for bug reports, feature requests, or improvements. 🙌
## License 📜
This project is licensed under the MIT License. See the
[LICENSE](LICENSE) file for details.
## Acknowledgments 💖
- [OpenCV](https://opencv.org/) for image processing 🖼️.
- [Tesseract OCR](https://github.com/tesseract-ocr/tesseract) for
Persian text extraction 🔍.
- [Pandas](https://pandas.pydata.org/) for data structuring 📚.