Proceedings of the IEEK Conference (대한전자공학회:학술대회논문집)

The Institute of Electronics and Information Engineers (대한전자공학회)
권호 표지

Optical Font Recognition For Printed Korean Characters Using Serif Pattern of Strokes

  • Kim, Soo-Hyung (Department of Computer Science, Chonnam National University) ;
  • Kim, Sam-Soo (Department of Computer Science, Chonnam National University) ;
  • Kwag, Hee-Kue (Department of Electronic Engineering and Computer Science, KAIST) ;
  • Lee, Guee-Sang (Department of Computer Science, Chonnam National University)
  • Published : 2002.07.01
Download PDF
⟨ Previous Next ⟩

Abstract

This paper introduces the problem of typeface classification of Hangul characters and proposes features for typeface classification among Serif and Sans-serif classes. Serif classes have a small decorative stroke around the beginning of vertical strokes, while Sans-serif classes have no serif. Therefore, the serif part is first segmented from the vertical strokes, and the direction of the serif is computed as the feature for Hangul typeface identification. To evaluate the performance of the proposed system, we used 3,000 characters extracted from Korean documents - 1,500 from Serif fonts, other 1,500 from Sans-serif fonts.

Keywords