HANDWRITING CHARACTER MODELING WITH IMPLICIT CURVES AND CLASSIFICATION

Article Name HANDWRITING CHARACTER MODELING WITH IMPLICIT CURVES AND CLASSIFICATION

Keywords TR El yazısı , karakter tanıma , kapalı cebirsel eğriler , modelleme , Bayes , yapay sinir ağları , uyarlamalı sinir-bulanık sınıflayıcı.

Keywords Handwriting , character recognition , implicit curves , modeling , Bayes , neural network , adaptive neuro-fuzzy classifier.

Article Summary In this study, the classification and modeling of handwriting characters by using the implicit curves were aimed. Also, the coefficients of eighth degree implicit equations were used for classification of handwritten digits. The Handwriting golden miners of gold miner games play online for adaptive neuro-fuzzy classifier by miner. Therefore, a variety of curve fitting methods were tested in the study. To be invariant, the normalization of the obtained coefficients was made according to only scaling and translation. Feature selection was also done with neuro-fuzzy classifier with linguistic hedges. In the study, the recognition rate of the method proposed with the Bayes and neural networks was measured by using entire and the certain part of MNIST database of handwritten digits. Recognition rate of 92.87 % was obtained, which in the study is not yet comparable to other methods. But it is clear that each character can be expressed with an equation by developing the method. By this way, the less memory using can be satisfied by storage of coefficients instead of the storage of image.

Download and Play Article Article Full Text

HANDWRITING CHARACTER MODELING WITH IMPLICIT CURVES AND CLASSIFICATION

Sigma Dergisi