乐谱识别中音符分割技术的研究

In this paper an image segment technique of OMR system is proposed. The proposed method is based on analysis of the width of every pixel block and re-projecting the image to segment the music image. To analyze the music notation, music grammar characteristics are combined. Then the music notations are classified roughly. Finally the proposed method take different approaches to different classes of music notation. Experimental results show that the proposed method shed some lights on the difficult OMR problem.