Title: Ethiopian banknote recognition using deep learning

Authors: Fitehalew Ashagrie Demilew; Gezahegn Mulusew Delele

Addresses: Debre Tabor University, Debre Tabor, South Gondar, Ethiopia ' Addis Ababa Science and Technology University, Addis Ababa, Ethiopia

Abstract: Paper currency recognition is one of the most challenging applications of pattern recognition. An automatic banknote recognition can be helpful in different aspects such as fraud detection and in automating services. The Ethiopian Government has introduced a new Ethiopian banknote denomination recently. Therefore, a recognition mechanism that can effectively identify the newly introduced banknotes should be proposed. The main challenges of banknote recognition are identifying fake and fraudulent banknotes. The paper proposes a deep learning-based banknote recognition system for the newly introduced Ethiopian banknotes. We have collected more than 5,000 images for the preparation of the dataset and each image has been resized into a size of 256*128. The dataset is divided into three sets with a ratio of 7:2:1 for training, testing and validation sets respectively. We have obtained an accuracy of 99.4% with a model loss of 0.01349 for the newly introduced Ethiopian banknotes.

Keywords: convolutional neural network; CNN; deep learning; Ethiopian banknote recognition; pre-processing.

DOI: 10.1504/IJITCC.2023.132845

International Journal of Information Technology, Communications and Convergence, 2023 Vol.4 No.2, pp.141 - 152

Accepted: 06 Feb 2023
Published online: 11 Aug 2023 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article