Description:

This dataset was developed to address the absence of publicly available datasets for Tunisian currency bills, which posed challenges in creating systems to assist the visually impaired in money identification. To overcome this gap, we built a comprehensive and diverse dataset of Tunisian currency bills, carefully capturing images under various conditions to simulate real-world scenarios.

Dataset Composition

The dataset includes high-resolution images of Tunisian banknotes across different denominations. We captured bills in various environments, different lighting conditions (from low light to bright daylight), and on diverse backgrounds. Additionally, the dataset includes manipulated images of bills (e.g., folded, crumpled) to replicate real-world usage.

Download Dataset

Image Variations

Lighting Conditions: Low, medium, and high light levels.

Background Types: Plain, textured, and cluttered.

Augmentation Techniques: Includes rotated, folded, and partially occluded bills to simulate practical scenarios.

Currency Denominations

The dataset features all common denominations of Tunisian bills, including:

5 Dinar

10 Dinar

20 Dinar

50 Dinar

100 Dinar

Applications

This dataset is ideal for:

Developing money identification tools for visually impaired individuals.

Enhancing digital wallet systems with currency recognition.

Training models for ATMs and other financial systems to recognize and authenticate bills.

Counterfeit detection research and development.

Summary

The Tunisian Currency Dataset is a crucial resource for developers and researchers focusing on currency recognition, especially those working on inclusive and assistive technologies. The variability in the dataset ensures that models trained on it can perform reliably in real-world conditions, contributing to more accessible and secure financial tools.

This dataset is sourced from Kaggle.