The Design of a Special Language for the Description of Hangul Patterns

Abstract

문법적 모형인식 시스템의 일반적 개념이 첫장에 소개된다. 그 다음에 한글 문자의 구조적 특성이 논의 된다. 좌에서 우로 스캐닝하는 방법이 제시되고, 그 스캐닝으로 부터 X-Y 좌표들의 수열이 구해진다. 그 수열은 한글모형의 숫자화된 이미지가 된다.

한글의 기본문자는 그 수열로 부터 연결된 점들을 모으는 알고리즘을 사용함으로써 추출된다. 그리고 그 기본문자 이미지로 부터 木構造가 구성된다. 그 木構造는 한글 모형을 기술하는 木構造문법에 적합한지 안한지 시험을 받게 된다. 정확한 木構造는 한글모형을 인식하는 木構造자동?館캇藪? 의해서 인식이 된다. 한글을 위한 木構造 문법과 木構造 자동인식기가 제시 되었다.

General concept of syntactic patten recognition system is introduced in the first chapter. Then, The structural characteristics of hangul characters are discussed. The left-right scanning method is presented. The sequences of X-Y coordinates are obtained from the scanning. They are the digitized images of hangul patterns. Basic characters are extracted by using the algorithm that collects the connected points from the sequences. From the basic character images the tree is constructed. Then, the tree is tested whether it is adaptable to the tree grammar for the hangul patterns. The tree grammar for the hangul patterns is presented. The correct tree is recognized by using the tree automaton for the hangul patterns. The tree automaton for the hangul patterns is presented. The program for basic character extraction algorithm is presented.

General concept of syntactic patten recognition system is introduced in the first chapter. Then, The structural characteristics of hangul characters are discussed. The left-right scanning method is presented. The sequences of X-Y coordinates are obtained from the scanning. They are the digitized images of hangul patterns. Basic characters are extracted by using the algorithm that collects the connected points from the sequences. From the basic character images the tree is constructed. Then, the tree is tested whether it is adaptable to the tree grammar for the hangul patterns. The tree grammar for the hangul patterns is presented. The correct tree is recognized by using the tree automaton for the hangul patterns. The tree automaton for the hangul patterns is presented. The program for basic character extraction algorithm is presented.