作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2010, Vol. 36 ›› Issue (7): 170-172. doi: 10.3969/j.issn.1000-3428.2010.07.058

• 人工智能及识别技术 • 上一篇    下一篇

手写汉语拼音的融合识别系统

朱 萌1,2,刘长松1,2,陈御天1,2,邹燕明3   

  1. (1. 清华大学智能技术与系统国家重点实验室,北京 100084;2. 清华大学电子工程系清华信息科学与技术国家实验室,北京 100084; 3. 诺基亚北京研究院,北京 100176)
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2010-04-05 发布日期:2010-04-05

Combined Recognition System for Handwritten Pinyin

ZHU Meng1,2, LIU Chang-song1,2, CHEN Yu-tian1,2, ZOU Yan-ming3   

  1. (1. State Key Laboratory of Intelligent Technology and Systems, Tsinghua University, Beijing 100084; 2. Tsinghua National Laboratory for Information Science and Technology, Department of Electronic Engineering, Tsinghua University, Beijing 100084; 3. Nokia Research Center, Beijing 100176)
  • Received:1900-01-01 Revised:1900-01-01 Online:2010-04-05 Published:2010-04-05

摘要: 手写设备用户容易忘记特定中文单字写法,需要为其提供拼音输入法。采用分类器融合方式构筑拼音单词识别系统,通过隐马尔可夫模型分类器获得拼音单词的切分点,利用统计特征识别模块进行识别后融合,研究并改进拼音单词基线提取方法。实验结果表明,该方法对17 745个测试样本的识别率达91.37%。

关键词: 中文信息处理, 字符识别, 基线

Abstract: Handwritten device users are easy to forget how to write a certain Chinese character. It is necessary to provide Pinyin input method for them. This paper constructs a Pinyin word recognition system through classifier fusion style. It obtains the cutting point of Pinyin word by Hidden Markov Model(HMM) classifier, accomplishes after-recognition fusion by using recognition module for statistic characteristic, studies and improves the base line extraction method for Pinyin word. Experimental results show that this method can recognize 91.37% test samples from 17 745 ones.

Key words: Chinese information processing, character recognition, base line

中图分类号: