https://github.com/neonwatty/python-ocr-preprocessing
Learn essential pre-processing techniques for effective Optical Character Recognition (OCR) in Python, including denoising, deskewing, and binarization.
https://github.com/neonwatty/python-ocr-preprocessing
binarization denoising deskewing optical-character-recognition python
Last synced: 8 months ago
JSON representation
Learn essential pre-processing techniques for effective Optical Character Recognition (OCR) in Python, including denoising, deskewing, and binarization.
- Host: GitHub
- URL: https://github.com/neonwatty/python-ocr-preprocessing
- Owner: neonwatty
- Created: 2024-06-08T22:48:50.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2025-01-12T02:12:10.000Z (about 1 year ago)
- Last Synced: 2025-04-06T03:41:26.649Z (11 months ago)
- Topics: binarization, denoising, deskewing, optical-character-recognition, python
- Language: Jupyter Notebook
- Homepage:
- Size: 3.6 MB
- Stars: 3
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Learn OCR pre-processing: denoising, deskewing, binarization, and more in Python
Explore the commonly overlooked pre-processing steps that help make Optical Character Recognition (OCR) models work properly in practice.
This repository contains code, a walkthrough notebook (`ocr_preprocessing_walkthrough.ipynb`), and streamlit demo app for playing around with common ocr pre-processing steps, and seeing their resulting effects on ocr quality.
All processing - from the various pre-processing steps to the ocr itself (here using the popular / classic [tesseract](https://github.com/tesseract-ocr/tesseract) model - are performed locally.
## Installation instructions
To create a handy tool for your own memes pull the repo and install the requirements file
```python
pip install -r requirements.txt
```
## Starting the streamlit app
Start the streamlit app by pasting the following in your terminal
```python
python -m streamlit run ocr/app.py
```
## Ocr your own images
Note: you can drag and drop any desired image directly into the streamlit app, and play around with how pre-processing steps effect the final ocr output.