摘要
该文介绍了维吾尔文的特点及维吾尔文字符识别系统 ,针对维吾尔文的连体结构 ,重点讨论了解决过程中的技术难点 .其中利用投影分离出连体段中的字母 ,采用边切分边识别的方法 ,对文本图像进行了切分 ,分类 ,提取外围特征 ,并通过样张的训练 ,使维吾尔文字符的识别获得了较满意的结果 .
This article introduces the application of the Uyghur sentence characteristic and Uyghur character recognition system to the connected structure of Uyghur. Authors discuss especially those aspects that solve the technological difficult points. The system utilizes projection to isolate the letters of the connected character field. Then the method is proposed, which recognises the characters when segmenting. By segmetation, classification,extraction of outer featuresof Uyghur text image, and the training of samples this character recognition algorithm can get satisfying results.
出处
《计算机学报》
EI
CSCD
北大核心
2004年第11期1480-1484,共5页
Chinese Journal of Computers
