Joint Banknote Recognition and Counterfeit Detection Using Explainable Artificial Intelligence
<p>Grad-CAM results for some example banknotes. The most left column to the third column show visible, infrared transmission, and infrared reflection images of banknotes. The forth column and the most right column show Grad-CAM results for banknote recognition and counterfeit detection, reflectively.</p> "> Figure 2
<p>Possible banknote directions.</p> "> Figure 3
<p>Different modality images. Leftmost column shows visible, center column shows infrared transmission, and rightmost column shows infrared reflection images.</p> "> Figure 4
<p>Joint banknote recognition and counterfeit detection system.</p> "> Figure 5
<p>pGrad-Cam flow.</p> "> Figure 6
<p>Average batch losses of convergence graphs.</p> ">
Abstract
:1. Introduction
2. Related Works
2.1. Sequential Banknote Recognition and Counterfeit Detection System
2.2. Grad-CAM
3. Contributions
4. Methods
4.1. Joint Banknote Recognition and Counterfeit Detection System
4.2. Explainable Artificial Intelligence
5. Experimental Results
6. Conclusions
Author Contributions
Funding
Acknowledgments
Conflicts of Interest
References
- Woo Lee, J.; Hong, H.; Wan Kim, K.; Ryoung Park, K. A Survey on Banknote Recognition Methods by Various Sensors. Sensors 2017, 17, 313. [Google Scholar] [CrossRef] [PubMed]
- Park, Y.; Kwon, S.; Pham, T.; Park, K.; Jeong, D.; Yoon, S. A high performance banknote recognition system based on a one-dimensional visible light line sensor. Sensors 2015, 15, 14093–14115. [Google Scholar] [CrossRef] [PubMed]
- Pham, T.D.; Nguyen, D.T.; Park, C.; Park, K.R. Deep Learning-Based Multinational Banknote Type and Fitness Classification with the Combined Images by Visible-Light Reflection and Infrared-Light Transmission Image Sensors. Sensors 2019, 19, 792. [Google Scholar] [CrossRef] [PubMed]
- Lee, S.H.; Lee, H.Y. Counterfeit Bill Detection Algorithm using Deep Learning. Int. J. Appl. Eng. Res. 2018, 13, 304–310. [Google Scholar]
- Pham, T.; Lee, D.; Park, K. Multi-national banknote classification based on visible-light line sensor and convolutional neural network. Sensors 2017, 17, 1595. [Google Scholar] [CrossRef] [PubMed]
- Sarfraz, M. An intelligent paper currency recognition system. Procedia Comput. Sci. 2015, 65, 538–545. [Google Scholar] [CrossRef]
- Krizhevsky, A.; Sutskever, I.; Hinton, G.E. Imagenet classification with deep convolutional neural networks. In Advances in Neural Information Processing Systems; The MIT Press: Cambridge, MA, USA, 2012; pp. 1097–1105. [Google Scholar]
- Szegedy, C.; Liu, W.; Jia, Y.; Sermanet, P.; Reed, S.; Anguelov, D.; Erhan, D.; Vanhoucke, V.; Rabinovich, A. Going deeper with convolutions. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Boston, MA, USA, 7–12 June 2015; pp. 1–9. [Google Scholar]
- Iandola, F.; Moskewicz, M.; Karayev, S.; Girshick, R.; Darrell, T.; Keutzer, K. Densenet: Implementing efficient convnet descriptor pyramids. arXiv 2014, arXiv:1404.1869. [Google Scholar]
- Howard, A.G.; Zhu, M.; Chen, B.; Kalenichenko, D.; Wang, W.; Weyand, T.; Andreetto, M.; Adam, H. Mobilenets: Efficient convolutional neural networks for mobile vision applications. arXiv 2017, arXiv:1704.04861. [Google Scholar]
- Ren, Y. Banknotes Recognition in Real Time Using ANN. Ph.D. Thesis, Auckland University of Technology, Auckland, New Zealand, 2017. [Google Scholar]
- Zhang, Q. Currency Recognition Using Deep Learning. Ph.D. Thesis, Auckland University of Technology, Auckland, New Zealand, 2018. [Google Scholar]
- Navya Krishna, G.; Sai Pooja, G.; Naga Sri Ram, B.; Yamini Radha, V.; Rajarajeswari, P. Recognition of fake currency note using convolutional neural networks. Int. J. Innov. Technol. Explor. Eng. 2019, 8, 58–63. [Google Scholar]
- Ba, J.; Caruana, R. Do deep nets really need to be deep? In Advances in Neural Information Processing Systems; The MIT Press: Cambridge, MA, USA, 2014; pp. 2654–2662. [Google Scholar]
- Samek, W.; Wiegand, T.; Müller, K.R. Explainable artificial intelligence: Understanding, visualizing and interpreting deep learning models. arXiv 2017, arXiv:1708.08296. [Google Scholar]
- Selvaraju, R.R.; Cogswell, M.; Das, A.; Vedantam, R.; Parikh, D.; Batra, D. Grad-CAM: Visual explanations from deep networks via gradient-based localization. In Proceedings of the IEEE International Conference on Computer Vision, Venice, Italy, 22–29 October 2017; pp. 618–626. [Google Scholar]
- Andreotti, F.; Phan, H.; De Vos, M. Visualising convolutional neural network decisions in automatic sleep scoring. In Proceedings of the Joint Workshop on Artificial Intelligence in Health (AIH) 2018, Stockholm, Sweden, 13–14 July 2018; pp. 70–81. [Google Scholar]
- Woo, S.; Park, J.; Lee, J.Y.; So Kweon, I. CBAM: Convolutional block attention module. In Proceedings of the European Conference on Computer Vision (ECCV), Munich, Germany, 8–14 September 2018; pp. 3–19. [Google Scholar]
- Graziani, M.; Andrearczyk, V.; Müller, H. Visual Interpretability for Patch-Based Classification of Breast Cancer Histopathology Images. 2018. Available online: https://openreview.net/forum?id=S1PTal9sz (accessed on 19 August 2019).
- He, K.; Zhang, X.; Ren, S.; Sun, J. Deep residual learning for image recognition. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, NV, USA, 26 June–1 July 2016; pp. 770–778. [Google Scholar]
- Liu, W.; Anguelov, D.; Erhan, D.; Szegedy, C.; Reed, S.; Fu, C.Y.; Berg, A.C. SSD: Single shot multibox detector. In European Conference on Computer Vision; Springer: Amsterdam, The Netherlands, 2016; pp. 21–37. [Google Scholar]
- Alshayeji, M.H.; Al-Rousan, M.; Hassoun, D.T. Detection method for counterfeit currency based on bit-plane slicing technique. Int. J. Multimed. Ubiquitous Eng. 2015, 10, 225–242. [Google Scholar] [CrossRef]
- Bhavani, R.; Karthikeyan, A. A novel method for counterfeit banknote detection. Int. J. Comput. Sci. Eng. 2014, 2, 165–167. [Google Scholar]
- Ambadiyil, S.; Reddy, T.; Teja, B.; Pillai, V. Banknote authentication using normalized cross correlation method. Discovery 2015, 44, 166–172. [Google Scholar]
- Lamsal, S.; Shakya, A. Counterfeit paper banknote identification based on color and texture. In Proceedings of the IOE Graduate Conference, Lalitpur, Nepal, 11–12 December 2015; pp. 160–168. [Google Scholar]
- Ching, T.; Himmelstein, D.S.; Beaulieu-Jones, B.K.; Kalinin, A.A.; Do, B.T.; Way, G.P.; Ferrero, E.; Agapow, P.M.; Zietz, M.; Hoffman, M.M.; et al. Opportunities and obstacles for deep learning in biology and medicine. J. R. Soc. Interface 2018, 15. [Google Scholar] [CrossRef] [PubMed]
- Han, S.S.; Kim, M.S.; Lim, W.; Park, G.H.; Park, I.; Chang, S.E. Classification of the clinical images for benign and malignant cutaneous tumors using a deep learning algorithm. J. Investig. Dermatol. 2018, 138, 1529–1538. [Google Scholar] [CrossRef] [PubMed]
- Martinel, N.; Foresti, G.L.; Micheloni, C. Wide-slice residual networks for food recognition. In Proceedings of the 2018 IEEE Winter Conference on Applications of Computer Vision (WACV), Lake Tahoe, NV, USA, 12–15 March 2018; pp. 567–576. [Google Scholar]
- Ioffe, S.; Szegedy, C. Batch normalization: Accelerating deep network training by reducing internal covariate shift. arXiv 2015, arXiv:1502.03167. [Google Scholar]
- Maas, A.L.; Hannun, A.Y.; Ng, A.Y. Rectifier nonlinearities improve neural network acoustic models. In Proceedings of the 2013 International Conference on Machine Learning (ICML), Atlanta GA, USA, 16–21 June 2013; Volume 30, p. 3. [Google Scholar]
- Sivadas, S.; Wu, Z.; Bin, M. Investigation of parametric rectified linear units for noise robust speech recognition. In Proceedings of the Sixteenth Annual Conference of the International Speech Communication Association, Dresden, Germany, 6–10 September 2015. [Google Scholar]
- Clevert, D.A.; Unterthiner, T.; Hochreiter, S. Fast and accurate deep network learning by exponential linear units (elus). arXiv 2015, arXiv:1511.07289. [Google Scholar]
- Bianco, S.; Cadene, R.; Celona, L.; Napoletano, P. Benchmark analysis of representative deep neural network architectures. IEEE Access 2018, 6, 64270–64277. [Google Scholar] [CrossRef]
- Goodfellow, I.; Bengio, Y.; Courville, A. Deep Learning; MIT Press: Cambridge, MA, USA, 2016; Available online: http://www.deeplearningbook.org (accessed on 19 August 2019).
- Srivastava, N.; Hinton, G.; Krizhevsky, A.; Sutskever, I.; Salakhutdinov, R. Dropout: A simple way to prevent neural networks from overfitting. J. Mach. Learn. Res. 2014, 15, 1929–1958. [Google Scholar]
- Wu, H.; Gu, X. Towards dropout training for convolutional neural networks. Neural Netw. 2015, 71, 1–10. [Google Scholar] [CrossRef] [Green Version]
- Kingma, D.P.; Ba, J. Adam: A method for stochastic optimization. arXiv 2014, arXiv:1412.6980. [Google Scholar]
Methods | Pros | Cons |
---|---|---|
Sequential method | - Relatively easy to train | - Relatively long inference time |
Joint method using CNN for image classification | - | - Extremely slow - Relatively difficult to train |
Proposed method | - Fast inference time | - Relatively difficult to train |
Nation | Denomination | The Number of Dataset | ||||||
---|---|---|---|---|---|---|---|---|
Train | Validation | Test | ||||||
Series | Type | Genuine | Counterfeit | Genuine | Counterfeit | Genuine | Counterfeit | |
EUR | First | 5 EUR | 3266 | 8 | 182 | 1 | 182 | 1 |
10 EUR | 744 | 76 | 42 | 2 | 42 | 2 | ||
20 EUR | 634 | 140 | 36 | 3 | 36 | 3 | ||
50 EUR | 864 | 536 | 49 | 8 | 49 | 8 | ||
100 EUR | 3110 | 216 | 173 | 3 | 173 | 3 | ||
200 EUR | 824 | 1364 | 46 | 20 | 46 | 20 | ||
500 EUR | 967 | 644 | 54 | 9 | 54 | 9 | ||
Second | 5 EUR | 1535 | 112 | 86 | 2 | 86 | 2 | |
10 EUR | 2304 | 40 | 129 | 1 | 129 | 1 | ||
20 EUR | 1371 | 268 | 77 | 4 | 77 | 4 | ||
50 EUR | 2702 | 72 | 151 | 1 | 151 | 1 | ||
Total | 18,321 | 3476 | 1025 | 54 | 1025 | 54 | ||
USD | 1 USD | 1735 | 0 | 97 | 0 | 97 | 0 | |
2 USD | 3780 | 0 | 210 | 0 | 210 | 0 | ||
5 USD | 2574 | 0 | 143 | 0 | 143 | 0 | ||
10 USD | 5298 | 1664 | 295 | 24 | 295 | 24 | ||
20 USD | 8369 | 5436 | 466 | 76 | 466 | 76 | ||
50 USD | 8263 | 340 | 460 | 5 | 460 | 5 | ||
100 USD | 3564 | 80 | 198 | 2 | 198 | 2 | ||
Total | 33,583 | 7520 | 1869 | 107 | 1869 | 107 |
Nation | Dataset | Number of Banknotes (Counterfeit) | Number of Well-Classified (Counterfeit) | Accuracy (%) | ||
---|---|---|---|---|---|---|
Sequential Method | Joint GoogleNet | Proposed Method | ||||
EUR | Train and validation | 22,876 (3530) | 22,876 (3530) | 22,876/22,876 (100) | 22,876/22,876 (100) | 22,876/22,876 (100) |
Test | 1079 (54) | 1079 (54) | 1079/1079 (100) | 1079/1079 (100) | 1079/1079 (100) | |
USD | Train and validation | 43,079 (7627) | 43,079 (7627) | 43,079/43,079 (100) | 43,079/43,079 (100) | 43,079/43,079 (100) |
Test | 1976 (107) | 1976 (107) | 1976/1976 (100) | 1976/1976 (100) | 1976/1976 (100) |
Model | Processing Time on Average (variance) | |||
---|---|---|---|---|
Preprocessing | Banknote Recognition | Counterfeit Detection | Total | |
Sequential | 3.57 ms (0.46) | 4.18 ms (0.86) | 3.94 ms (0.69) | 11.69 ms (3.02) |
Joint GoogleNet | 3.53 ms (0.12) | 947.12 ms (2817.44) | 950.65 ms (2804.65) | |
Proposed | 3.73 ms (0.49) | 4.36 ms (0.72) | 8.09 ms (1.18) |
Banknote | Input Images | Banknote Recognition | Counterfeit Detection | ||||
---|---|---|---|---|---|---|---|
Visible | Infrared Transmission | Infrared Reflection | Grad-CAM | pGrad-CAM | Grad-CAM | pGrad-CAM | |
20 EUR first series | |||||||
20 EUR second series | |||||||
200 EUR | |||||||
500 EUR | |||||||
1 USD | |||||||
2 USD | |||||||
50 USD | |||||||
100 USD |
Banknote | Method | Input Example Images | Explainable Artificial Intelligence | ||||
---|---|---|---|---|---|---|---|
Visible | Infrared Transmission | Infrared Reflection | Banknote Denomination | Banknote Direction | Counterfeit Detection | ||
20 EUR first series | Grad-CAM | ||||||
pGrad-CAM | |||||||
100 EUR | Grad-CAM | ||||||
pGrad-CAM | |||||||
10 USD | Grad-CAM | ||||||
pGrad-CAM | |||||||
20 USD | Grad-CAM | ||||||
pGrad-CAM |
© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
Share and Cite
Han, M.; Kim, J. Joint Banknote Recognition and Counterfeit Detection Using Explainable Artificial Intelligence. Sensors 2019, 19, 3607. https://doi.org/10.3390/s19163607
Han M, Kim J. Joint Banknote Recognition and Counterfeit Detection Using Explainable Artificial Intelligence. Sensors. 2019; 19(16):3607. https://doi.org/10.3390/s19163607
Chicago/Turabian StyleHan, Miseon, and Jeongtae Kim. 2019. "Joint Banknote Recognition and Counterfeit Detection Using Explainable Artificial Intelligence" Sensors 19, no. 16: 3607. https://doi.org/10.3390/s19163607