Dataset of Peruvian Banknotes

Recognizing and categorizing banknotes is a crucial task, especially for individuals with visual impairments. It plays a vital role in assisting them with everyday financial transactions, such as making purchases or accessing their workplaces or educational institutions. The primary objectives for creating this dataset were as follows:

1. To compile a dataset containing both the old and new Peruvian currency.
2. To establish a dataset exclusively for Peruvian currency.
3. To ensure that the dataset comprises high-quality images.

We utilized the rear cameras of mobile phones to capture images of Peruvian banknotes. The processed image dataset comprises a total of 9,315 images, with 6,568 for training, 2,486 for validation, and 261 for testing. This processed dataset encompasses 16 distinct categories, including the obverse and reverse sides of 10, 20, 50, and 100 old Peruvian soles, as well as the obverse and reverse sides of 10, 20, 50, and 100 new Peruvian soles. The images were taken under various backgrounds and lighting conditions, including cluttered backgrounds and images of folded banknotes. Researchers specializing in banknote detection and recognition will find this dataset highly valuable.