USE OF A PRINCIPAL COMPONENT ANALYSIS FOR THE RECOGNITION OF THE GRAPHIC OBJECTS
Abstract
The simple procedure of the formation of the optimum filters of signs on the basis of principal component is proposed. The use of optimum filters for the solution of the problem of recognition
makes it possible to build the trivial neuron networks (without the learning and the hidden layers) in the form of the classifiers of signs.
About the Authors
A. Gubin
Санкт-Петербургский государственный университет телекоммуникаций им. проф. М.А. Бонч-Бруевича
Russian Federation
Russian Federation
V. Litvinov
Санкт-Петербургский государственный университет телекоммуникаций им. проф. М.А. Бонч-Бруевича
Russian Federation
Russian Federation
F. Filippov
Санкт-Петербургский государственный университет телекоммуникаций им. проф. М.А. Бонч-Бруевича
Russian Federation
Russian Federation
References
1. Krizhevsky A., Sutskever I., Hinton G. E. ImageNet classification with deep convolutional neural networks // Advances in Neural Information Processing Systems 25, 2012, pp. 1106–1114.
2. Convolutional Neural Networks for Visual Recognition, Stanford CS class CS231n. URL: http://cs231n.github.io/convolutional-networks/.
Review
For citations:
Gubin A., Litvinov V., Filippov F. USE OF A PRINCIPAL COMPONENT ANALYSIS FOR THE RECOGNITION OF THE GRAPHIC OBJECTS. Proceedings of Telecommunication Universities. 2016;2(3):27-31. (In Russ.)
Views:
206
Email this article 