Handwritten Chinese Character Recognition Based onPrimitive and Fuzzy Features via the SEART Neural Net Model

Authors:
Hahn-Ming Lee;Chung-Chieh Sheu;Jyh-Ming Chen
Affiliations:
Department of Electronic Engineering, National Taiwan University of Science and Technology, Taipei, Taiwan. E-mail: hmlee@et.ntust.edu.tw;Department of Electronic Engineering, National Taiwan University of Science and Technology, Taipei, Taiwan. E-mail: hmlee@et.ntust.edu.tw;Department of Electronic Engineering, National Taiwan University of Science and Technology, Taipei, Taiwan. E-mail: hmlee@et.ntust.edu.tw
Venue:
Applied Intelligence
Year:
1998

Citing 10
Cited 1

A modified fast parallel algorithm for thinning digital patterns

Pattern Recognition Letters
On machine recognition of hand-painted Chinese characters by feature relaxation

Pattern Recognition
Neural network models for pattern recognition and associative memory

Neural Networks
A Nearest Hyperrectangle Learning Method

Machine Learning
A new approach to stroke and feature point extraction in Chinese character recognition

Pattern Recognition Letters
On speeding candidate selection in handprinted Chinese character recognition

Pattern Recognition
Fuzzy ART: Fast stable learning and categorization of analog patterns by an adaptive resonance system

Neural Networks
False stroke detection and elimination for character recognition

Pattern Recognition Letters
Neural networks at work

IEEE Spectrum
A Bayesian neural network for separating similar complex handwritten Chinese characters

Pattern Recognition Letters

Diagnosis Based on Genetic Algorithms and Fuzzy Logic in NPPs

Proceedings of the 14th International conference on Industrial and engineering applications of artificial intelligence and expert systems: engineering of intelligent systems

Quantified Score

Hi-index	0.00

Visualization

Abstract

A handwritten Chinese character recognition method based onprimitive andcompound fuzzy features using the SEART neural network model isproposed. The primitive features are extracted in local andglobal view. Since handwritten Chinese characters vary a greatdeal, the fuzzy concept is used to extract the compound featuresin structural view. We combine the two categories of featuresand use a fast classifier, called the Supervised Extended ART(SEART) neural network model, to recognize handwritten Chinesecharacters. The SEART classifier has excellent performance, isfast, and has good generalization and exception handling abilities incomplex problems. Using the fuzzy set theory in featureextraction and the neural network model as a classifier ishelpful for reducing distortions, noise and variations. Inspite of the poor thinning, a 90.24%recognition rate on average for the 605test character categories was obtained. The database used isCCL/HCCR3 (provided by CCL, ITRI, Taiwan). The experiment notonly confirms the feasibility of the proposed system, but alsosuggests that applying the fuzzy set theory and neural networks torecognition of handwritten Chinese characters is an efficientand promising approach.