Image Dataset Splitter

📂 A tool for splitting image datasets into training and testing sets for classification tasks. Useful for preparing data for machine learning models.

Features

• Splits image datasets into training and testing subsets.
• Supports organizing images by classification labels.
• Generates a YAML configuration file for dataset management.

Installation

1. Clone the repository:

   git clone https://github.com/Gulciha-n/image-classification-data-split.git
   cd image-classification-data-split
   

2. Install required packages:

   pip install pyyaml

Usage

1. Prepare your dataset:

   Ensure your source directory is structured as follows:

   source/
   └── train/
       ├── belirsiz/
       ├── sifir/
       ├── on/
       ├── yirmi/
       ├── otuz/
       ├── kirk/
       ├── elli/
       ├── altmis/
       ├── yetmis/
       ├── seksen/
       ├── doksan/
       └── yuz/
   

   Here, source is your dataset directory, and each subdirectory (e.g., belirsiz, sifir, etc.) contains images corresponding to the classification labels.

2. Update the script with your paths:

   Edit the main.py file and update the source_dir and target_dir variables with your directory paths.

   source_dir = r"PATH_TO_YOUR_SOURCE_DIRECTORY"  # source directory
   target_dir = r"PATH_TO_YOUR_TARGET_DIRECTORY"  # target directory

3. Run the script:

   Execute the following command to run the script:

   python main.py

   This will organize and split your image dataset into training and testing sets, and generate a dataset.yaml file in the target directory.

   After running the script, your target directory structure will be:

   train_test_images/
   ├── train/
   │   ├── belirsiz/
   │   ├── sifir/
   │   ├── on/
   │   ├── yirmi/
   │   ├── otuz/
   │   ├── kirk/
   │   ├── elli/
   │   ├── altmis/
   │   ├── yetmis/
   │   ├── seksen/
   │   ├── doksan/
   │   └── yuz/
   └── test/
       ├── belirsiz/
       ├── sifir/
       ├── on/
       ├── yirmi/
       ├── otuz/
       ├── kirk/
       ├── elli/
       ├── altmis/
       ├── yetmis/
       ├── seksen/
       ├── doksan/
       └── yuz/
   

   A dataset.yaml file will also be created in the train_test_images directory.

Configuration

• source_dir: Directory where your image data is located.
• target_dir: Directory where the split data will be saved.
• labels: List of classification labels used for organizing images.

License

This project is licensed under the MIT License