A Python script for generating an OCR dataset for Geez scripts including Amharic and Tigrinya.
# OCR Dataset Generation for Geez Scripts (Amharic and Tigrinya)
This is a Python script for generating an OCR dataset for Geez scripts including Amharic and Tigrinya. The dataset consists of PNG images of each character generated using a set of fonts and background/foreground colors. The purpose of this project is to provide a simple and customizable way of generating a dataset for training optical character recognition (OCR) models for the Amharic and Tigrinya languages.
## Uses and Advantages
The dataset generated by this script can be used to train OCR models for recognizing Geez script characters in Amharic and Tigrinya languages. The advantages of this project include:
* Customizability: The user can choose the fonts, font sizes, background and foreground colors for the generated characters.
* Large Dataset: The script can generate a large number of character images, which can be used to train OCR models with high accuracy.
* Easy to Use: The script can be run from the command line with arguments for customizing the dataset generation process.
## Methodology
The script generates PNG images of each character in the Amharic and Tigrinya languages using a set of fonts, font sizes, and background/foreground colors. The characters are centered in the image and randomly positioned. The generated images are saved in a specified directory with a unique filename for each image.
## Contribution
Contributions to this project are welcome. You can contribute by creating pull requests or reporting issues. Some potential areas for contribution include:
* Adding support for additional fonts
* Adding support for additional languages
* Improving the script's performance or efficiency
## How to Use
To use the script, you must have Python 3 and the Pillow library installed. You can install Pillow using pip with the command pip install pillow.
To generate the dataset, run the generate_dataset.py script with the following command line arguments:
```
python generate_data …