This project implements a simple image processing pipeline in Python using Pillow and NumPy. It takes a folder of images, performs a series of transformations, and saves the processed images to a new directory.
- Resizing: Resizes images to a specified square size while maintaining aspect ratio.
- Padding: Adds padding to resized images to ensure they are perfectly square.
- Batch Processing: Processes all images within a specified input directory.
- Unique Output Directory: Creates a new output directory with a unique timestamp for each run.
-
Clone the repository:
git clone [https://dictionnaire.reverso.net/francais-definition/non+valide](https://dictionnaire.reverso.net/francais-definition/non+valide) cd python-image-processing -
Install dependencies:
pip install -r requirements.txt
-
Organize your images:
Place the images you want to process in the
input_imagesdirectory. -
Run the pipeline:
python main.py
This will process all images in the
input_imagesdirectory and save the processed images to a new directory within thedatasetfolder.
You can customize the pipeline by modifying the main.py file:
destination_size: Change this variable to adjust the desired output size for the images.
This project uses Ruff for linting and formatting, and pre-commit for managing pre-commit hooks.
-
Install development dependencies:
pip install -r requirements.txt
-
Install pre-commit hooks:
pre-commit install