2D Geometric Shapes Dataset

Description

The repository contains a Python script for generating a 2D geometric shapes dataset, together with the dataset itself. The dataset includes 16 different geometric shapes, each randomly oriented and positioned within 224 × 224 pixel images.

Included Shapes

• Circle
• Semicircle
• Ellipse
• Triangle
• Square
• Rectangle
• Parallelogram
• Rhombus
• Trapezoid
• Kite shape
• Pentagon
• Hexagon
• Heptagon
• Octagon
• Nonagon
• Decagon

Dataset Structure

• shapes_dataset/: Contains subdirectories for each shape, each with 100 PNG images.
  • circle/
    • circle_1.png
    • circle_2.png
    • ...
  • semicircle/
  • ...
• Total Images: 1,600 (100 images × 16 shapes)

Installation

To generate the dataset in your own environment, clone this repository and run the script:

git clone https://github.com/Khalid1G/shapes-dataset.git
cd shapes-dataset
pip install pillow
python generate_shapes_dataset.py

License

The dataset is released under the MIT License.

Usage

The dataset can be used for various purposes, such as training machine learning models for shape classification or object detection tasks. Images are saved in PNG format with transparent backgrounds, facilitating overlay on other images.

Contribution

Contributions to this repository are welcome. Some ideas include:

• Adding new shapes to the dataset
• Improving the script to generate more diverse shapes
• Creating additional scripts for preprocessing or augmenting the dataset
• Using the dataset in machine learning projects and sharing results

If you wish to contribute, open an issue or pull request.

Acknowledgements

The dataset was created by Khalid1G as a personal project. Shapes are generated using the Python Imaging Library (PIL) and saved as PNG images.

Contact

If you have any questions about the dataset, feel free to contact me via:

• khalid.boussaroual@gmail.com
• kboussaroual@insea.ac.ma
• LinkedIn
• Or open an issue in this repository.