Bangladeshi Banknote

Description

To develop and evaluate the banknote serial number identification system, I constructed a dataset of Bangladeshi banknote images. The images were captured using an 8-megapixel smartphone camera to reflect realistic acquisition conditions. Banknotes were collected from personal contacts, banking institutions, and retail stores, ensuring diversity in circulation quality and physical condition. The dataset contains 450 images captured under varying real-world conditions, including different illumination levels, viewing angles, backgrounds, and perspectives. It includes Bangladeshi banknotes of 1000, 500, 200, and 100 Taka notes. Each image contains a visible banknote serial number, which serves as the target for serial number detection and recognition. To improve the robustness and generalizability of the method, the dataset incorporates substantial variability in banknote appearance, including differences in color, wear and tear, folds, stains, partial occlusions, and imaging conditions. These characteristics closely resemble practical deployment scenarios and provide a challenging benchmark for evaluating banknote serial number detection and recognition algorithms.

Institutions

• Khulna University

  Khulna Division, Khulna

Categories

Licence