本文描述了一个基于HMM模型的联机汉字识别系统的设计思想与实现方法。系统以联机汉字的笔段序列作为观察序列,采用带有多跨越的模型结构消除自由书写汉字笔段序列的冗余与丢失问题。HMM模型的训练是本系统设计的一个重要问题,针对复杂HMM模型参数训练容易收敛于局部最小的情况,本文结合联机汉字识别的特点,提出了一种利用“引导模型”进行训练的改进方法,避免了训练过程收敛于局部最小点的发生。经过大量样本的训练,本系统对规范书写汉字和自由书写汉字均取得了比较令人满意的结果。
Abstract
This paper describes the design and implementation of an on-line Chinese Character recognition system , which is based on Hidden Markov Models1 The strokes of on-line Chinese character are regarded as the input observation sequence , and a multi-cross left-right model structure is employed in order to eliminate the influence caused by redundancy or loosing of strokes. The training of HMM models is also an important problem for this system , in order to avoid the training process falls into local minimum , an improved training approach is proposed. After sufficient training , this system gains an satisfying result for both ordinary writing characters and free-style writing characters.
关键词
隐含Markov模型 /
联机汉字识别
{{custom_keyword}} /
Key words
hidden Markov model /
on-line Chinese character recognition
{{custom_keyword}} /
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] Liu Jiafeng , Tang Jianglong ,Shu Wenhao. A Structure Similarity Analysis method for On-Line Recognition of Handwritten Chinese Characters. Journal of Harbin Institute of Technology ,1994 ,E - 1 (1) : 51 - 54
[2] Toru Wakahara , Hiroshi murase , Kazumi Odaka. On-line Handwritting Recognition. Proceedings of the IEEE ,1992 ,80 (7) :1181 - 1194
[3] 唐降龙,孙广玲,刘家锋. 一种笔段序列匹配联机汉字识别方法. 计算机研究与发展,1999 ,36 (12) : 1472 - 1476
[4] Mou-Yen Chen ,Amlan Kundu ,Jian Zhou. Off-Line Handwritten Word Recognition using a Hidden Markov Model Type Stochasic Network. IEEE Transactions on Pattern Analysis and Machine Intelligence ,1994 ,16 (5) :481 - 496
[5] Der-Sheng Lin ,Jin-Jang Leou ,A Genetic Algorithm Approach to Chinese Handwriting Normalization. IEEE Transaction on Systems ,Man and Cybernetics - Part B :CYBERNETICS ,1997 ,27 (6) :999 - 1006
{{custom_fnGroup.title_cn}}
脚注
{{custom_fn.content}}
基金
863高技术研究发展计划(863-306-ZD03-02-06)
{{custom_fund}}
PDF(149 KB)
