원문정보
The Music Score Recognition System of the Robust Music Symbols Distortion for Computer Games
초록
영어
Introduce the score-recognition method. E.g The note, beat and other symbols for the play music. Usually, the score-recognition is using the template matching or statistical model. This paper suggest the new method that is using the template matching and feature of pixel pattern for computer games. Typically, the morphology of general music symbol are independent on the change of font. For instance the head and line of note, staccato, crescendo, decrescendo and so on. But some music symbol aren’t keep their morphology. For instance the beat of the note, chord and beam note. This paper applying the template matching at the general music symbol and applying the feature of pixel pattern analysis method at the some music symbol. So this paper can get great accuracy of score-recognition. This paper can recognize the Whole, Half, Quarter, Eighth and Sixteenth note and rest, Seminote, Staccato, Crescendo, Decrescendo, Key signature and Clef
목차
1. 서론
2. 기존의 악보 인식 방법
2.1 선 추적과 템플릿 매칭을 이용한 악보 인식 기술
2.2 신경망을 이용한 기법
3. 제안하는 악보 검출 방법
3.1 악보 영상 전처리
3.2 오선 검출 단계
3.3 객체 검출 단계
4. 제안하는 악보 인식 방법
4.1 음표 인식
4.2 비 음표 인식 단계
4.3 Music XML
5. 시뮬레이션 결과본
6. 결론
감사의 글
참고문헌