Traffic Sign Images from Turkey

The dataset consists of 21,249 images, though not all are 32×32 pixels. In the accompanying notebook, all images are resized to 32×32 pixels for consistency. This preprocessing step ensures that the dataset is ready for training classification models. The diversity and quantity of the images make this dataset ideal for educational purposes and initial computer vision experiments.

Dataset Features:

Total Images: 21,249

Image Resolution: Various (resized to 32×32 pixels in the provided notebook)

Categories: Different types of traffic signs commonly found in Turkey

Data Augmentation: Applied to increase variation and improve model training

Acknowledgements

This dataset would not be possible without the support and contributions of the community. We extend our gratitude to all the individuals who have helped collect and annotate the images. If you use this dataset in your research or projects, please consider citing relevant past work and providing appropriate attributions.

Inspiration

Many students and beginners in computer vision undertake projects to build and refine their skills. This dataset aims to assist in understanding the basics of computer vision, offering a practical and manageable challenge. By working with this dataset, students can gain hands-on experience in image classification, data augmentation, and model training.

Additional Information

To further enhance your learning experience, the dataset includes a sample notebook demonstrating essential preprocessing steps, model training, and evaluation. This notebook serves as a guide to help you get started with the dataset and develop your computer vision project.

Use Cases:

Educational projects to understand computer vision fundamentals

Practice for data augmentation techniques

Training basic image classification models

Experimenting with preprocessing methods