DOI 10.17586/0021-3454-2021-64-3-202-207
UDC 004.89
DEVELOPMENT OF A CONVOLUTIONAL LAYER OF A DEEP NEURAL NETWORK FOR DETECTING DEFECTS IN ROLLED METAL
Murom Institute — Branch of Vladimir State University, Department of Computer-aided Designing System of Electronic Devices; Post-Graduate Student
A. L. Zhiznyakov
Murom Institute — Branch of Vladimir State University, Department of Computer-aided Designing System of Electronic Devices; Professor, Head of Department
D. V. Titov
Southwest State University, Department of Computer Engineering; Professor
K. V. Mortin
Murom Institute of Vladimir State University, Department of Software Engineering;
Read the full article
Abstract. A convolutional layer of a deep neural network, designed to determine defects in rolled metal products, is considered. To determine the defect, it is proposed to use algorithms for the segmentation of flaw detection images and several types of filtering within the convolutional layer. Filtering is based on the use of combined convolution algorithms with different initial masks. To minimize the error at the output of the convolutional layer, the GELU activation function is used. Results of the experiments are presented.
Keywords: flaw detection image, convolutional layer, filtering, deep neural network, convolution
References:
References:
- Barkhatov V.A. Russian Journal of Nondestructive Testing, 2006, no. 2, pp. 92–100.
- Barkhatov V.A. Russian Journal of Nondestructive Testing, 2006, no. 4, pp. 227–236.
- Barkhatov V.A. Russian Journal of Nondestructive Testing, 2009, no. 2, pp. 73–85.
- Gurvich А.К. Kontrol’. Diagnostika (Testing. Diagnostics), 2013, no. 10, pp. 68–70. (in Russ.)
- Krasnoshchekov A.A., Sobol B.V., Solov'ev A.N., Cherpakov A.V. Russian Journal of Nondestructive Testing, 2011, no. 6, pp. 412–419.
- Nikiforova T.V. Matematicheskiye metody raspoznavaniya obrazov (Mathematical Methods Pattern Recognition), 12th All-Russian Conference, Materials of reports, 2005, рр. 392–395. (in Russ.)
- Butenko V. V. Molodoy uchenyy (Young Scientist), 2015, no. 4, pp. 52–56. (in Russ.)
- Artemov A.A., Kavalerov M.V., Kuznetsov G.S. Vestnik PNIPU. Elektrotekhnika, informatsionnyye tekhnologii, sistemy upravleniya, 2011, no. 5. (in Russ.)
- Akimov A.V., Sirota A.A. Computer Optics, 2016, no. 6, pp. 899–906. (in Russ.)
- Viola P., Jones M. International Journal of Computer Vision, 2004, no. 2(57), pp. 137–154.
- Girshick R. International Conference on Computer Vision (ICC), 2015.
- Girshick R., Shaoqing R., Kaiming H. Neural Information Processing Systems (NIPS), 2015.
- Wang Y., Wang C., Zhan H., Yingbo G., Wei S. Remote Sensing, 2019, no. 11, pp. 531.
- Silver D., Huang A., Maddison C. et al. Nature, 2016, no. 529(7587), pp. 484–489.
- Wainberg M., Alipanahi B., Frey B.J. Journal of Machine Learning Research, 2016, no. 17(110), pp. 1–5.
