摘要
本文提出了一种在隐含马尔可夫模型(HMM)框架下建立的识别脱机手写汉字的方法,介绍了以HMM对脱机手写汉字进行建模、识别的整个过程,并给出了实验结果.对国标一级3755个汉字的识别率,在两种测试集上分别达到96.4%和91.5%.
In this paper, an efficient scheme for off-line handwritten Chinese character recognition in the framework of Hidden Markov Model(HMM) is proposed. The whole process of modeling and recognizing off-line handwritten Chinese character is presented. The experimental result is also given. The recognition rate for 3755 Chinese characters in the basic set of the national standard GB2312 achieves 96. 4% and 91.5% on two sources of samples respectively.
出处
《模式识别与人工智能》
EI
CSCD
北大核心
2002年第1期84-88,共5页
Pattern Recognition and Artificial Intelligence
