Development of a convolutional neural network model for the classification of Russian traffic signs

Authors: Matveev D.A., Petrunicheva A.S.
Published in issue: #9(50)/2020
DOI: 10.18698/2541-8009-2020-9-640
Category: Informatics, Computer Engineering and Control | Chapter: System Analysis, Control, and Information Processing, Statistics
Keywords: computer vision, convolutional neural network, neural network training, image classification problem, traffic sign classification, Russian Traffic Sign Dataset, Google Colaboratory, Keras
Published: 19.10.2020

The traffic sign classification system can be used both as an independent solution for image analysis tasks, and as part of a car computer vision system for solving problems of driver assistance and autonomous control. The paper considers some modern approaches in the field of image classification using convolutional neural networks, analyzes their typical architectures and features, describes the stages of developing and training a model of such a neural network. The presented results were obtained based on the use of the database of images of Russian traffic signs — Russian Traffic Sign Dataset (RTSD). The software implementation of the algorithms for training the neural network model and its analysis was carried out in Python using the Keras library on the public platform Google Colaboratory.

References

[1] Shruti J., Sudip P. Recent trends in image and signal processing in computer vision. Springer, 2020.

[2] Kamusoko C. Image classification. In: Remote sensing image classification in R. Springer, 2019, pp. 81–153.

[3] Russakovsky O., Deng J., Su H., et al. ImageNet large scale visual recognition challenge. Int. J. Comput. Vis., 2015, vol. 115, no. 3, pp. 211–252. DOI: https://doi.org/10.1007/s11263-015-0816-y

[4] Shadman S., Nazib A., Kabir J.A., et al. An overview of convolutional neural network. Preprints, 2018, art. 2018110546. DOI: https://doi.org/10.20944/preprints201811.0546.v1

[5] O’Shea K., Nash R. An introduction to convolutional neural networks. arxiv.org: website. URL: https://arxiv.org/abs/1511.08458v2 (accessed: 20.03.2020).

[6] LeCun Y., Bottou L., Bengio Y., et al. Gradient-based learning applied to document recognition. Proc. IEEE, 1998, vol. 86, no. 11, pp. 2278–2324. DOI: https://doi.org/10.1109/5.726791

[7] Krizhevsky A., Sutskever I., Hinton G.E., et al. Imagenet classification with deep convolutional neural networks. Proc. NIPS, 2012, pp. 1097–1105.

[8] Simonyan K., Zisserman A. Very deep convolutional networks for large-scale image recognition. arxiv.org: website. URL: https://arxiv.org/abs/1409.1556 (accessed: 20.03.2020).

[9] Szegedy Ch., Liu W., Jia Ya., et al. Going deeper with convolutions. IEEE CVPR, 2015. DOI: https://doi.org/10.1109/CVPR.2015.7298594

[10] He K., Zhang X., Ren Sh., et al. Deep residual learning for image recognition. IEEE CVPR, 2016, pp. 770–778. DOI: https://doi.org/10.1109/CVPR.2016.90

[11] Shakhuro V.I., Konushin A.S. Russian traffic sign images dataset. Komp’yuternaya optika [Computer Optics], 2016, vol. 40, no. 2, pp. 294–300. DOI: https://doi.org/10.18287/2412-6179-2016-40-2-294-300 (in Russ.).

[12] Kingma D.P., Ba J.L. Adam: a method for stochastic optimization. arxiv.org: website. URL: https://arxiv.org/abs/1412.6980 (accessed: 20.03.2020).

[13] Ajiboye A., Arshah A., Qin R., et al. Evaluating the effect of dataset size on predictive model using supervised learning technique. IJSECS, 2015, 1. pp. 74–84. DOI: https://doi.org/10.15282/ijsecs.1.2015.6.0006