Text Segmentation and Extraction from Images of Guqin Jianzi Pu
Changle Zhou · Mind and Computation · 2007
To transcribe Guqin Jianzi pu (reduced notation) is a highly-professional, time-consuming and laborious work. If to identify and transcribe characters in Jianzi pu can be pursued automatically by means of advanced artificial intelligence techniques, even if the work is subsidiary it will play an important role in the development of Guqin Jianzi pu transcription and contribute to protect and promote Guqin culture as well. This paper investigates the attributes of Guqin Jianzi pu and proposes an OCR method for segmentation and extraction notations from images of Jianzi pu, including algorithms and implementations of pretreatment of initial scanned images of Jianzi pu, segmentation and extraction of Jianzi pu from mixed notation of Jianzi pu and stuff notation and segmentation and extraction of characters from a single line Jianzi pu. This work orients to the particularities of Guqin Jianzi pu and performs universally, so to some extent it also enriches the research on Chinese character segmentation.