https://github.com/inquilabee/tablecv

TableCV: Table extraction from images made easy.
https://github.com/inquilabee/tablecv

opencv opencv-python opencv-table opencv-table-extraction python table table-extract table-extract-python table-extraction

Last synced: about 2 months ago
JSON representation

TableCV: Table extraction from images made easy.

• Host: GitHub
• URL: https://github.com/inquilabee/tablecv
• Owner: inquilabee
• License: mit
• Created: 2023-09-14T20:52:58.000Z (over 1 year ago)
• Default Branch: main
• Last Pushed: 2023-09-14T21:38:37.000Z (over 1 year ago)
• Last Synced: 2024-03-21T11:48:06.320Z (10 months ago)
• Topics: opencv, opencv-python, opencv-table, opencv-table-extraction, python, table, table-extract, table-extract-python, table-extraction
• Language: Python
• Homepage:
• Size: 107 KB
• Stars: 3
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# TableCV

**TableCV** is a Python package designed to extract tables from images. It offers two approaches for extracting tables, allowing you to choose the one that best suits your needs.

## Installation

You can easily install **TableCV** using pip:

```bash

pip install tablecv

```

## Usage

### Approach 1 (using PaddleOCR)

**TableCV** offers a straightforward method to extract tables using PaddleOCR. This approach returns a pandas DataFrame object:

```python

from tablecv import extract_table

# Replace "image_path" with the path to your image

print(extract_table(image_path="your_image.png"))

```

### Approach 2 (OCR with Your Preferred Tool)

If you prefer using a different OCR tool like EasyOCR, KerasOCR, or any other OCR solution, you can still use **TableCV**. First, perform OCR on your image using your chosen tool. The OCR results should be structured as a list of tuples, each containing a bounding box and corresponding text:

```python

# List of tuples: (bounding box as (x, y, w, h), text)

ocr_results = [

    ((1, 2, 3, 4), "a"),

    ((4, 5, 6, 7), "b"),

    # Add more tuples as needed

]

```

After obtaining your OCR results, you can extract tables from them using **TableCV**:

```python

from tablecv import extract_table_from_ocr

# Replace "ocr_results" with your OCR results list

print(extract_table_from_ocr(ocr_results))

```

With these two approaches, **TableCV** provides flexibility for table extraction from images, whether you prefer using PaddleOCR or another OCR tool of your choice.