About some approaches to problem of metals and alloys microstructures classification based on neural network technologies

DOI

10.2991/aime-18.2018.56How to use a DOI?

Keywords

functional materials; strength properties; complex estimation; grain-phase structure; problem of segmentation; classification of microstructures; neural network technologies

Abstract

Two approaches to the decision of the problem of of metal alloys microstructures classification using neural network technologies are considered. Characteristics of the existing methods of recognition of grains circuits and phases of difficult microstructures are selected and described. A certain sequence of application of methods of the segmentation problem decision and classification by means of the developed neuronets is offered. Two alternative candidate solutions of the classification problem depending on adequate accuracy and volume of a set of input data are offered. Learning database for neural net with providing examples that show differences of same picture on various scales is described. The mathematical feature of operation of each of methods is described. Examples of architecture of a neuronet and samples of input data are given.

Copyright

© 2018, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - R. Klestov
AU  - A. Klyuev
AU  - V. Stolbov
PY  - 2018/04
DA  - 2018/04
TI  - About some approaches to problem of metals and alloys microstructures classification based on neural network technologies
BT  - Proceedings of the International Conference "Actual Issues of Mechanical Engineering" (AIME 2018)
PB  - Atlantis Press
SP  - 292
EP  - 296
SN  - 2352-5401
UR  - https://doi.org/10.2991/aime-18.2018.56
DO  - 10.2991/aime-18.2018.56
ID  - Klestov2018/04
ER  -