USAGE OF CONVOLUTIONAL NEURAL NETWORKS FOR IMAGE CLASSIFICATION
https://doi.org/10.57070/2304-4497-2023-1(43)-39-49
Abstract
The paper discusses the structure of a convolutional neural network and the mathematical methods used to calculate its values. The main components of the network that affect the result are given: convolutional layers with a mask as the basis of the data network, a core for reading the data network, steps and additions for adjusting the reading accuracy, subsampling layers for generalizing data. The history of the development of convolutional neural networks with examples of their architecture and parameters used is shown on the example of networks LeNet, AlexNet, VGG and ResNet. A comparison of the accuracy of pattern recognition for different architectures is shown. The concept of transfer learning is described.
About the Authors
Alexander G. Bychkov
Siberian State Industrial University
Russian Federation
Russian Federation
Postgraduate of Department of Applied Information Technologies and Programming
Tamara V. Kiseleva
Siberian State Industrial Universit
Dr. Sci. (Eng.), Prof. of the Department of Applied Information Technologies and Programming
Elena V. Maslova
Siberian State Industrial University
Cand. Sci. (Eng.), Assist. Prof. of the Department of Applied Information Technologies and Programming
References
1. Komashinskii V.I. Neural networks and their application in control and communication systems. Moscow: Hotline-Telecom, 2002, 94 p. (In Russ).
2. Kiseleva T.V., Maslova E.V., Bychkov A.G. Machine learning for solving image recognition problems. In: Proceedings of the 50th International Conference on Information Technologies in Science, Education and Management. Gurzuf, 2021, pp. 19–24. (In Russ).
3. Kiseleva T.V., Maslova E.V., Bychkov A.G. Machine learning in image recognition tasks. Informatting and communication. 2021, no. 8, pp. 15–19. (In Russ).
4. Bychkov A.G., Kiseleva T.V., Maslova E.V. Methods of optimization of neural network elements by the example of a perceptron. Management systems and information technologies. 2022, no. 1 (87), pp. 4–8. (In Russ). https: //doi.org/10.36622/VSTU.2022.88.1.001/
5. Chervyakov N.I. Application of neural networks for forecasting tasks and problems of identification of forecasting models. Neurocomputers: development and application. 2003, no. 10, pp. 11–14. (In Russ).
6. Kruglov V.V. Artificial neural networks. Theory and practice. Moscow: Hotline-Telecom, 2001, 382 p. (In Russ).
7. Khaikin S. Neural networks. Full course. Moscow: Williams Publishing House, 2006, 1104 p. (In Russ).
8. Nikolenko S. Deep learning. Sankt-Peterburg: Piter, 2018, 480 p. (In Russ).
9. Bychkov A.G., Kiseleva T.V., Maslova E.V. Using segmentation in convolutional neural networks to improve accuracy. Management systems and information technologies. 2022, no. 3 (89), pp. 7–10. (In Russ). https: //doi.org/10.36622/VSTU.2022.88.3.003/
10. LeCun Y., Boser B., Denker J.S., Henderson D., Howard R.E., Hubbard W., Jackel L.D. Backpropagation applied to handwritten zip code recognition. Neural Computation. 1989, vol. 1 (4), pp. 541–551.
11. LeCun Y. Generalization and network design strategies. In: Technical Report CRG-TR-89-4 Department of Computer Science, University of Toronto, 1989, pp. 1276–1289.
12. LeCun Y., Boser B., Denker J.S., Henderson D., Howard R.E., Hubbard W., Jackel L.D. Handwritten digit recognition with a back-propagation network. In: Advances in Neural Information Processing Systems 2 (NIPS 89). 1990, pp. 1008–1019.
13. LeCun Y., Bottou L., Bengio Y., Haffner P. Gradient-based learning applied to document recognition. Proceedings of the IEEE. 1998, vol. 86 (11), pp. 2278–2324.
14. Chellapilla K., Puri S., Simard P. High Performance Convolutional Neural Networks for Document Processing. In: Tenth International Workshop on Frontiers in Handwriting Recognition. Guy L. ed. Suvisoft. 2006, pp. 2877–2889.
15. Krizhevsky A., Sutskever I., Hinton G.E. Imagenet classification with deep convolutional neural networks. In: Advances in neural information processing systems. 2012, pp. 1097–1105.
16. Shelhammer L.E., Darrell T. Fully convolu-tional networks for semantic segmentation. In: The IEEE Conf. On Computer Vision and Pattern Recognition (CVPR). 2015, pp. 3431–3440.
17. Badrinarayanan V., Kendall A., Cipolla R. SegNet: A deep convolutional encoder-decoder architecture for image segmentation. ArXiv. 2015, vol. 1511, article 00561. https://doi.org/10.48550/arXiv.1511.00561/
18. Ronneberger O., Fischer P., Brox T. U-net: convolutional networks for biomedical image segmentation. In: Proc. Med. Image Comput. Comput.-Assisted Intevention. 2015, pp. 234–241.
19. Iglovikov V., Shvets A. TernausNet: U-net with vgg11 encoder pre-trained on imagenet for image segmentation. ArXiv. 2018, vol. 1801, article 05746.
20. Santos J., Ferro E., Orozco J., Cayssials R. A heuristic approach to the multitask-multiprocessor assignment problem using the empty-slots method and rate monotonic scheduling. J. of Real-Time Systems. 1997, vol. 13 (2), pp. 167–199.
Review
For citations:
Bychkov A., Kiseleva T., Maslova E. USAGE OF CONVOLUTIONAL NEURAL NETWORKS FOR IMAGE CLASSIFICATION. Bulletin of the Siberian State Industrial University. 2023;(1):39-49. (In Russ.) https://doi.org/10.57070/2304-4497-2023-1(43)-39-49
Views:
22
Email this article 