Turkish Lira Banknote Classification using Transfer Learning and Deep Learning

Abstract

With the increasing exchange of foreign currencies due to globalization, there is a need for systems that can recognize and validate multiple currencies in real time. Such systems facilitate smooth international transactions and support the finance sector in dealing with diverse currencies. This study focuses on classifying Turkish banknotes using deep learning models. The dataset comprises 6901 images of six different denominations (5 TL, 10 TL, 20 TL, 50 TL, 100 TL, and 200 TL) under various conditions, such as flat, angled, curved, and bent. The proposed model imple ments pre-trained models, including VGG16, VGG19, DenseNet121, DenseNet169, DenseNet201, MobileNet, and MobileNetV2, to classify the images. Different image sizes (50x50, 100x100, 150x150, and 200x200) and optimizers (SGD, RMSprop, Adam, Adamax, etc.) were tested to determine the most effective combinations. The best result was achieved with DenseNet201 with an image size of 200 and the SGDoptimizer, achieving an accuracy of 98.84% in 12 epochs. Smaller image sizes (50x50) resulted in reduced performance for all models. In addition, models such as DenseNet169 and DenseNet121 also demonstrated high performance; however, MobileNetV2 struggled with smaller images.

Keywords

• DenseNet201
• Optimizer
• Banknote
• Convolution
• Accuracy

References

1. Baek, S., Choi, E., Baek, Y., & Lee, C. (2018). Detection of counterfeit banknotes using multispectral images. Digital Signal Processing, 78, 294—304. https://doi.Org/10.1016/j.dsp.2018.03.015 google scholar
2. Baltaci, F. (2020). Turkish lira banknote dataset. Retrieved January 2, 2024, from https://www.kaggle.com/datasets/baltacifatih/ turkish- lira- banknote- dataset google scholar
3. Baykal, G., Demir, U., Shyti, I., & Unal, G. (2018). Turkish lira banknotes classification using deep convolutional neural networks. In 2018 26th Signal Processing and Communications Applications Conference (SIU) (pp. 1-4). IEEE. google scholar
4. Filter, J. (2022). Split-folders - PyPI. Retrieved January 15, 2024. from https://pypi.org/project/split-folders/ google scholar
5. Foody, G. M. (2023). Challenges in the real world use of classification accuracy metrics: From recall and precision to the Matthews correlation coefficient. PLOS ONE, 18(10), e0291908. https://doi.org/10.1371/journal.pone.0291908 google scholar
6. Galeana Pérez, D., & Bayro Corrochano, E. (2018). Recognition system for euro and Mexican banknotes based on deep learning with real scene images. Computación y Sistemas, 22(4). https://doi.org/10.13053/cys-22-4-3079 google scholar
7. iyikesici, B., & Ergelebi, E. (2023). An efficient deep learning architecture for Turkish lira recognition and counterfeit detection. Turkish Journal of Electrical Engineering and Computer Sciences, 31(3), 678-692. https://doi.org/10.55730/1300-0632.4009 google scholar
8. Keras Team. (2023a). Keras application. Keras. Retrieved January 15, 2024, from https://keras.io/api/applications/ google scholar

9. Keras Team. (2023b). Optimizers. Keras. Retrieved January 17, 2024, from https://keras.io/api/optimizers/ google scholar
10. Khashman, A., Ahmed, W., & Mammadli, S. (2019). Banknote issuing country identification using image processing and neural networks. In Proceedings (pp. 746-753). https://doi.org/10.1007/978-3-030-04164-9_98 google scholar
11. Khashman, A., & Sekeroglu, B. (2005). Multi-banknote identification using a single neural network. In Proceedings (pp. 123-129). google scholar
12. Khashman, A., Sekeroglu, B., & Dimililer, K. (2005). Deformed banknote identification using pattern averaging and neural networks. In Proceedings of the 4th WSEAS International Conference on Computational Intelligence, Man-Machine Systems and Cybernetics (pp. 233-237). google scholar
13. Linkon, A. H. M., Labib, M. M., Bappy, F. H., Sarker, S., Jannat, M.-E., & Islam, M. S. (2020). Deep learning approach combining lightweight CNN architecture with transfer learning: An automatic approach for the detection and recognition of Bangladeshi banknotes. In 2020 11th International Conference on Electrical and Computer Engineering (ICECE) (pp. 214-217). IEEE. https://doi.org/10.1109/icece51571.2020. 9393113 google scholar
14. Mittal, S., & Mittal, S. (2018). Indian banknote recognition using convolutional neural network. In 2018 3rd International Conference On Internet of Things: Smart Innovation and Usages (IoT-SIU) (pp. 1-6). IEEE. https://doi.org/10.1109/iot- siu.2018.8519888 google scholar
15. Pachón, C. G., Ballesteros, D. M., & Renza, D. (2023). An efficient deep learning model using network pruning for fake banknote recognition. Expert Systems with Applications, 233, 120961. https://doi.org/10.1016/j.eswa.2023.120961 google scholar
16. Prakash, H., Yadav, A., Ushashree, P., Jha, C., Sah, G. K., & Naik, A. (2023). Deep learning approaches for automated detection of fake Indian banknotes. In 2023 IEEE International Conference on Integrated Circuits and Communication Systems (ICICACS) (pp. 1-5). IEEE. https://doi.org/10.1109/icicacs57338.2023.10100265 google scholar
17. Sahin, O. (2018). GitHub - Ozgurshn/TurkishBanknoteDataset. Retrieved January 2, 2024, from https://github.com/ozgurshn/ TurkishBanknoteDataset. google scholar
18. Veeramsetty, V., Singal, G., & Badal, T. (2020). CoinNet: Platform independent application to recognize Indian currency notes using deep learning techniques. Multimedia Tools and Applications, 79(31-32), 22569-22594. https://doi.org/10.1007/s11042-020-09031-0 google scholar
19. Wang, L., Zhang, Y., Lanchi, X., Zhang, X., Guang, X., Li, Z., Li, Z., Shi, G., Hu, X., & Zhang, N. (2022). Automated detection and classification of counterfeit banknotes using quantitative features captured by spectral-domain optical coherence tomography. Science & Justice, 62(5), 624-631.https://doi.org/10.1016/j.scijus.2022.09.004. google scholar