Realization of information technology of character recognition based on competing cellular automata
| Autor: | Vladimir Zhikharevich, Ivan Myroniv, S. E. Ostapov |
|---|---|
| Rok vydání: | 2017 |
| Předmět: |
Theoretical computer science
Computer science business.industry Process (engineering) Applied Mathematics Mechanical Engineering media_common.quotation_subject Realization (linguistics) Energy Engineering and Power Technology Industrial and Manufacturing Engineering Cellular automaton Computer Science Applications Mobile automaton Set (abstract data type) Control and Systems Engineering Management of Technology and Innovation Quality (business) Performance indicator Artificial intelligence Simplicity Electrical and Electronic Engineering business media_common |
| Zdroj: | Eastern-European Journal of Enterprise Technologies. 3:18-24 |
| ISSN: | 1729-4061 1729-3774 |
| DOI: | 10.15587/1729-4061.2017.103550 |
| Popis: | We examined the possibility of applying cellular automata to solve the problem on recognition of text characters. For this purpose, we introduced the notion of competing cellular automata and developed algorithms of their functioning and interaction. In order to implement the proposed algorithms, we created modeling software. It allowed us to evaluate effectiveness of the cellular-automaton algorithms to conduct experiments on text character recognition using the English alphabet and to demonstrate a number of advantages in comparison with other methods. We investigated a description of the solution to the problem on selecting the structural attributes in the images of text characters, which directly affect the quality of recognition. In the present work, it is proposed to use a set of cellular automata constructed on the diagrams of states of Moore and Mealy machines for each type of cellular automata, to determine the end points, junctions and cycles in characters. We considered operation of the modeling program. The advantages of the given technology are the simplicity of rules of interaction, easy parallelization of the process of recognition, the possibility of recognition of distorted and partially overlapping characters. We compared performance quality and efficiency of the commercially available system ABBYY FineReader, which has demonstrated high performance indicators of the developed recognition technology |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|