Related Experiment Video
Updated: Mar 12, 2026
13:19
Deep Neural Networks for Image-Based Dietary Assessment
Published on: March 13, 2021
10.1K
Kazakh banknote image dataset
Ualikhan Sadyk1, Makhambet Yerzhan2, Cemil Turan1
1Department of Computer Science, SDU University, Abylaikhan St. 1/1, Kaskelen, Kazakhstan.
Data in Brief
|March 11, 2026
Summary
This study introduces a new image dataset of Kazakhstani tenge banknotes for computer vision research. The dataset aids in developing and testing banknote recognition systems under diverse conditions.
Area of Science:
- Computer Vision
- Pattern Recognition
- Machine Learning
Background:
- Developing robust automated systems for currency recognition is crucial.
- Existing datasets may lack diversity in real-world conditions like varied backgrounds and lighting.
- Kazakhstani tenge recognition requires specialized datasets for accurate classification and detection.
Purpose of the Study:
- To present a comprehensive image dataset of Kazakhstani banknotes.
- To facilitate research in computer vision and pattern recognition tasks.
- To support the development of advanced banknote identification algorithms.
Main Methods:
- Collected and curated a dataset of Kazakhstani tenge banknotes.
- Photographed banknotes across seven denominations (500-20,000 Tenge) and a mixed category.
- Ensured image diversity with varied backgrounds, lighting, and multiple/single note configurations.
Main Results:
- Dataset includes over 100 images per denomination and 1000+ for mixed categories.
- Specific subsets address old/new designs and single/multiple note scenarios for the 5000 Tenge.
- Images capture realistic variations in background and illumination.
Conclusions:
- The dataset provides a valuable resource for training and evaluating banknote recognition models.
- Enables research on model robustness against environmental variations.
- Supports transfer learning, bias studies, and comparative evaluations in currency recognition.
