NSTU-BDTAKA: NSTU Bangladeshi Paper Currency Dataset

The "NSTU-BDTAKA" dataset has been carefully organized to facilitate two primary tasks: the identification and detection of taka, with a particular emphasis on Bangladeshi paper money. A hierarchical folder structure inside the "Detect" folder divides data into "train," "test," and "validation" subsets for the Taka detection subset, with each including "images" and "labels" subfolders. Annotation files in YOLO format are kept in the "labels" folder, which guarantees accurate object localization for YOLOv5-based object identification models. The dataset is divided into "train," "test," and "validation" subsets in the taka recognition subset, with distinct class folders for different amounts of money. These class folders include images of particular denominations, which simplifies the dataset for training recognition models and allows for precise currency classification based just on image data. This meticulous structure supports the development, training, and evaluation of models, making the "NSTU-BDTAKA" dataset a valuable resource for researchers and practitioners in computer vision and currency analysis.