基于特征点的汉字字体识别研究

王恺; 靳简明; 史广顺; 王庆人

doi:10.3724/SP.J.1146.2006.01129

基于特征点的汉字字体识别研究

doi: 10.3724/SP.J.1146.2006.01129

计量
- 文章访问数: 3359
- HTML全文浏览量: 92
- PDF下载量: 1256
- 被引次数: 0
出版历程
- 收稿日期: 2006-07-31
- 修回日期: 2007-05-07
- 刊出日期: 2008-02-19

Chinese Font Recognition Based on Feature Point

摘要

摘要: 该文提出了整体分析法和个体分析法的概念，并在分析它们各自适用范围的基础上，指出个体分析法更适合于解决印刷体汉字字体识别。在此基础上，提出一种基于特征点的个体分析法来解决汉字字体识别问题，与以往方法相比，该方法具有3个优点：识别可信度可控；处理速度快；适用于多语混排情况。实验结果表明，该方法有效解决了印刷体汉字字体识别问题，其性能大大优于以往方法。
- 字体识别; 光学字符识别; 特征点
Abstract: Global analysis method and individual analysis method are proposed in this paper. By analyzing their traits, it is concluded that individual analysis method is more suitable for machine-printed Chinese font recognition. A feature point based individual analysis method is proposed to resolve Chinese font recognition problem. Compared with previous methods, there are mainly three advantages: The recognition reliability is controllable; the processing speed is fast; it is suitable for multi-lingual document image processing. Experimental results show that the proposed method is more effective than previous methods.

HTML全文

参考文献(1)

曾理, 唐远炎, 陈廷槐. 基于多尺度小波纹理分析的文字种类自动识别. 计算机学报, 2000, 23(7): 699-704.Zeng Li, Tang Yuan-yan and Chen Ting-huai. Multi-scalewavelet texture-based script identification method. ChineseJournal of Computers, 2000, 23(7): 699-704.[2]Zhu Yong, Tan Tieniu, and Wang Yunhong. Fontrecognition based on global texture analysis[J].IEEETransactions on Pattern Analysis and Machine Intelligence.2001, 23(10):1192-1200[3]王洪, 汪同庆, 刘建胜, 朱永权, 黄甫征声. 基于小波包纹理分析的字体识别方法. 光电工程, 2002, 29(S1): 62-65.Wang Hong, Wang Tong-qing, Liu Jian-sheng, ZhuYong-quan,and Huangfu Zheng-sheng. Method for characterfont recognition based on wavelet package texture analysis.Opto-electronic Engineering, 2002, 29(S1): 62-65.[4]杨志华, 齐东旭, 杨力华, 吴力军. 基于经验模式分解的汉字字体识别方法. 软件学报, 2005, 16(8): 1438-1444.Yang Zhi-hua, Qi Dong-xu, Yang Li-hua, and Wu Li-jun. AChinese font recognition method based on empirical modedecomposition. Journal of Software, 2005, 16(8): 1438-1444.[5]Yang Zhihua, Yang Li-hua, and Suen Ching Y. A newmethod of recognizing Chinese fonts. Proceedings of the 2005Eight International Conference on Document Analysis andRecognition, Seoul, Korea, 29 August - 1 September 2005:962-966.[6]陈力, 丁晓青. 基于小波特征的单字符汉字字体识别. 电子学报, 2004, 32(2): 177-180.Chen Li and Ding Xiao-qing. Font recognition of singleChinese character based on wavelet feature. Acta ElectronicaSinica, 2004, 32(2): 177-180.[7]王恺, 王庆人. 中英文混合文章识别研究. 软件学报, 2005,16(5): 786-798.Wang Kai and Wang Qing-ren. Research on Chinese/Englishmixed document recognition. Journal of Software, 2005,16(5): 786-798.[8]张炘中, 闫昌德, 刘秀英. 汉字识别的特征点法及其一种应用.中文信息学报, 1987, 11(3): 13-19.Zhang Xin-zhong, Yan Chang-de and Liu Xiu-ying. Amethod of chinese recognition based on characteristic dotMatching. Journal of Chinese Information Processing, 1987,11(3): 13-19.[9]刘维一, 盛益强, 乔辉, 方志良. 特征点匹配法实现汽车牌照的快速识别. 光电子激光, 2002, 13(3): 274-276.Liu Wei-yi, Sheng Yi-qiang, Qiao Hui, and Fang Zhi-liang.Characteristic dot matching method of realizing rapidlyrecognition of the number plate. Journal of Optoelectronics.Laser, 2002, 13(3): 274-276.[10]邢向华, 顾国华. 基于模板匹配和特征点匹配相结合的快速车牌识别方法. 光电子技术, 2003, 23(4): 268-270.Xing Xiang-hua and Gu Guo-hua. Method of quicklyrecognizing vehicle plate based on pattern matching andcharacteristic dot matching. Optoelectronic Technology, 2003,23(4): 268-270.[11]靳简明. 数学公式图像处理研究. [博士论文], 南开大学, 2003.Jin Jian-ming. Research on typeset mathematical expressionimage processing. [Ph.D.dissertation], Nankai University, 2003.

施引文献

资源附件(0)

访问统计