计算机光学乐谱识别技术  

Application and Research on Optical Music Recognition

作　　者：刘晓翔[1] 张树生[1] 王贺[1] 朱玉璋[2] 

机构地区：[1]西北工业大学,西安710072 [2]西安音乐学院,西安710061

出　　处：《计算机工程》2003年第2期14-15,27,共3页Computer Engineering

基　　金：西北工业大学博士论文创新基金资助项目

摘　　要：计算机光学乐谱识别是计算机技术在音乐领域的发展和应用，主要利用图像处理、模式识别、文档图像分析等相关技术，把乐谱图像转化成通用的数字音乐格式。OMR实现传统乐谱数字化，在计算机音乐、计算机辅助音乐教学、数字音乐图书馆等众多领域有着广泛的应用前景。OMR包括乐谱图像预处理、谱线检测与删除、原始音符对象识别和特征音符对象解释与重组等主要过程，音符对象的识别、解释与重组是其中的难点和关键。The paper introduces the technique of optical music recognition (OMR) which suggests the automatic recognition of music notation by computer. An OMR system uses a scanned image of a printed score as input, while its output could be a text-to-audio transcribed file, to provide user with editing or hearing functionalities of the recognition results. Subsequent processing on this output can provide a wide variety of applications, such as reprinting, computer aid music teaching and digital music library, etc. Then, this paper discusses four key areas of OMR techniques: image pre-processing; staff line detection and removing; primitive musical objects recognition; feature musical objects reasoning and reconstruction.

关 键 词：计算机 光学乐谱识别 模式识别 音乐教学 计算机辅助教学 数学音乐图书馆 文档图像分析 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术] J6[自动化与计算机技术—计算机科学与技术]