RECOGNITION SYSTEM OF WRITING SYMBOLS BY NEURAL NETWORK

Abstract

Neural systems find an increasing number of spheres where they can be applied. It can be use from helping to protect information on mobile devices and to create any prediction of any social, economic or natural processes. The article will try to answer on question of how neural system can determine handwritten symbols. During the analysis of this issue it will be determined how it is possible to teach the system to recognize symbols and how to improve the operation of this algorithm by self-learning algorithm of neural system. Also we will try answer on question of work neural network with incorrect input data and perspective of neural network. In whole the article need to answer the question of neural network mission and do we need to use it or we have another alternatives.

Authors and Affiliations

Д. Г. Соколенко, Я. І. Корнага

Keywords

Related Articles

ANALYSIS OF TECHNICAL AND ECONOMIC INDICATORS OF PLOWING LAND PLOTS OF TRACTORS OF VARIOUS CAPACITY

The work is devoted to the analysis of efficiency of soil tillage by tractors of different productive capacity. The resulted technique of definition of economically proved optimal quantity of agricultural machinery and t...

RADIO-ELECTRONIC DEVICE FOR FORMING DETERMINISTIC CHAOS SIGNALS BASED ON THE DUFFING-HOLMES NON-LINEAR SYSTEM

The paper presents results of theoretical research and simulation of the radio-electronic device for forming the deterministic chaos signals. The Duffing-Holmes non-linear system, which is basic in the deterministic chao...

TO THE PROJECT OF ECOLOGICALLY CLEAN AND SAFE VITAL ACTIVITY ON BARREN AND POLLUTED TERRITORIES

Decision of modern ecological problems, including tasks of recultivation and rehabilitation of the land resources, by practical application of organo-mineral composites (ОМC) and selective sorbents (ZlP) on the basis of...

PRIMARY PROCESSING AND IDENTIFICATION OF BIOMETRIC INFORMATION FROM THE HEART RATE MONITOR

In the article was considered the primary processing of pulsograms with the using of hardware and software filtration. It is used for further identification and comparison with pulsograms characterizing the pathology by...

OPTIMIZATION OF THE PARAMETERS OF THE MOBILE ROBOT WITH GAS GENERATOR OF ENERGY

The article describes a fundamentally new design of a mobile robot with a gas energy storage and a mathematical model for optimizing its parameters. The optimization criteria for the design and technological parameters t...

Download PDF file
  • EP ID EP579953
  • DOI -
  • Views 46
  • Downloads 0

How To Cite

Д. Г. Соколенко, Я. І. Корнага (2018). RECOGNITION SYSTEM OF WRITING SYMBOLS BY NEURAL NETWORK. Вчені записки Таврійського національного університету імені В. І. Вернадського. Серія: Технічні науки, 29(5), 56-59. https://europub.co.uk/articles/-A-579953