Development of Recognition System for Template-marked Digits in Billets
| Authors: | Ko Jihun, Agency for Defense Development, Korea, Republic of
Lee YoungKow, Pohang Univ. of Science and Technology, Korea, Republic of
Choi SungHoo, Pohang Univ. of Science and Technology, Korea, Republic of
Kim Sang Woo, Pohang Univ. of Science and Technology, Korea, Republic of
Lee JongHak, POSCO Technology Laboratory, Korea, Republic of |
|---|
| Topic: | 1.1 Modelling, Identification & Signal Processing |
|---|
| Session: | Image Processing and Biomedical Applications |
|---|
| Keywords: | Pattern recognitionIdentificationImage processing |
|---|
Abstract
This paper presents a new solution to recognize template-markeddigits in billets. Since these digits can be easily blurred, asystematic approach to preprocessing and a recognition algorithmwith better generalization property are required. In our approach,the preprocessing methods using the local threshold and skewcorrection extract features with good quality from the capturedbillet images and multi-class support vector machine (SVM) basedalgorithms selecting the final class among the potential classesare applied. The experimental results show that the proposedapproach has higher recognition rate than that of the previousmethod using the k-nearest neighborhood(K-NN) technique fortemplate-marked digits in billets.
