Bangladeshi Counterfeit Currency Image Dataset
Description
Description: This dataset contains 1,286 labeled images of Bangladeshi banknotes collected for counterfeit currency detection. It includes both genuine and counterfeit notes of the 500 BDT and 1000 BDT denominations, captured under real-world conditions using standard cameras. Dataset Content: Total Images: 1,286 Genuine Notes: 1,199 Counterfeit Notes: 87 Denomination-wise Distribution 500 BDT: 829 genuine, 32 counterfeit (861 total) 1000 BDT: 370 genuine, 55 counterfeit (425 total) Purpose: The dataset is intended to support research in image-based counterfeit detection, binary classification, and anomaly detection for Bangladeshi currency notes.
Files
Steps to reproduce
The dataset can be reproduced by downloading and extracting the images from Mendeley Data, organizing them by denomination and class labels (genuine or counterfeit), resizing images to a fixed resolution, and splitting them into training and testing sets. The images can then be used directly for binary classification or anomaly detection experiments using standard machine learning or deep learning models.
Institutions
- Daffodil International University