A genetic algorithm approach to Chinese handwriting normalization

Author

Lin, Der-Sheng ; Leou, Jin-Jang

Author_Institution

Res. & Dev. Centre, Matsushita Electr. Ind. Co. Ltd., Taipei, Taiwan

Volume

27

Issue

6

fYear

1997

fDate

12/1/1997 12:00:00 AM

Firstpage

999

Lastpage

1007

Abstract

Normalization can be used to absorb writing variations and distortions, simplify the recognition processing steps, and improve the recognition rate of a Chinese handwriting recognition system. In this study, a genetic algorithm approach to Chinese handwriting normalization is proposed. In the proposed approach, a generalized normalization transform is defined as a linearly weighted combination of several normalization transforms and then genetic algorithms (GA´s) are used to determine the optimal set of weighting coefficients. Here the fitness function contains three proposed features representing the characteristics of Chinese characters, namely, stroke density variation (SDV), character area coverage (CAC), and centroid offset (CO). Experimental results show the feasibility of the proposed approach

Keywords

genetic algorithms; handwriting recognition; Chinese; centroid offset; character area coverage; fitness function; genetic algorithm; handwriting recognition; normalization; stroke density variation; Computer science; Councils; Equations; Feature extraction; Genetic algorithms; Handwriting recognition; Image segmentation; Linearity; Research and development; Writing;

fLanguage

English

Journal_Title

Systems, Man, and Cybernetics, Part B: Cybernetics, IEEE Transactions on

Publisher

ieee

ISSN

1083-4419

Type

jour

DOI

10.1109/3477.650059

Filename

650059