UDC 004.93’1
NEURAL NETWORK FOR IMAGE RECOGNITION OF MAGNETOOPTIC MATERIALS DOMAIN STRUCTURES
A. V. Bragin, senior lecturer of Radio engineering department, N. P. Ogarev Mordovia State University; This email address is being protected from spambots. You need JavaScript enabled to view it.
R. R. Navletov, post-graduate student, N.P. Ogarev Mordovia State University; This email address is being protected from spambots. You need JavaScript enabled to view it.
D. V. Pyanzin, PhD (technical sciences) associate professor of Radio engineering department, N.P. Ogarev Mordovia State University; This email address is being protected from spambots. You need JavaScript enabled to view it.
Computer vision systems are often used to study the structure of materials allowing to improve the quality of the results and reduce the time spent. The basis of these systems is neural network architecture. The article describes a neural network for image recognition labyrinthine domain structure. Such a structure is formed in magneto-optical materials being necessary in optoelectronic devices, spin electronics, magnetophotonics.
The aim is to develop a neural network of direct distribution for image recognition labyrinthine domain structure and classification of shape (round, elliptical, dumbbell, bandpass and branched) objects. Registration of domain structures of images is a magneto-optical unit, developed at the Department of Radio Engineering of National Research Mordovia State University [6.9]. As a result two-layer neural network of direct distribution is built and trained, the software for magnetooptical system is developed.
Key words: neural network, training, labyrinthine domain structure, informative signs, magneto-optical materials.