MST在手写汉字切分中的应用-CiteSeerX.PDFVIP

ISSN 1000-9825, CODEN RUXUEW E-mail: jos@ Journal of Software, Vol.17, No.3, March 2006, pp .403−409 http://www.j DOI: 10. 1360/jos170403 Tel/Fax : +86-10 © 2006 by Journal of Software. All rights reserved. ∗ MST 在手写汉字切分中的应用 1+ 1 2 韩 勇 , 须 德 , 戴国忠 1(北京交通大学 计算机与信息技术学院,北京 100044) 2( 中国科学院 软件研究所 人机交互技术与智能信息处理实验室,北京 100080) Using MST in Handwritten Chinese Characters Segmentation 1+ 1 2 HAN Yong , XU De , DAI Guo-Zhong 1(School of Computer Information Technology, Beijing Jiaotong University, Beijing 100044, China) 2(Laboratory of Human-Computer Interaction and Intelligent Information Processing, Institute of Software, The Chinese Academy of Sciences, Beijing 100080, China) + Corresponding author: Phn: +86-10 E-mail: hanyong_sc@, Han Y, Xu D, Dai GZ. Using MST in handwritten Chinese characters segmentation. Journal of Software, 2006,17(3):403−409. /1000-9825/17/403.htm Abstract: Handwritten Chinese characters segmentation is to process strokes based on its spatial relations to form character elements for recognition. This paper introduces a method to segment Chinese characters according to the topological relations of Chinese component and minimal span tree of strokes. The experiment shows that this new method can achieve good performance. The accuracy of segmentation is over 91.6%. Key words: character segmentation; handwritten Chinese character; component structure; online recognition 摘 要: 手写汉字切分是根据输入笔迹的空间位置关系进行汉字部件的合并切分,形成完整的汉字笔划以便 进行识别处理.综合利用了汉字部件的结构位置关系和笔划的空间位置关系,根据笔划的最小生成树(minimal spanning tree, 简称MST)对联机连续手写输入汉字进行切分,取得了较好的切分结果.切分的准确率超过91.6%. 关键