【作者】 周纬； 陈良育； 曾振柄；

【Author】 Zhou Wei,Chen Liangyu,Zeng Zhenbin (College of East China Normal University,Shanghai 200062,China)

【机构】 华东师范大学上海市高可信计算重点实验室；

【摘要】 在分析藏文的几何结构的基础上,本文提出了一种基于几何形状分析的藏文识别方法。文章通过对藏文的辅音和元音符号的几何特征提取,达到学习的目的,再根据印刷体藏文的组成结构特征,对其进行切分,进而分别对切分结果进行识别,将藏文字符转译为拉丁文,使得即使不认识藏文的人也能"读"藏文。实验表明,该方法的正确率迭89%以上。更多还原

【Abstract】 This paper presented an approach to recognize the Tibetan by analyzing the structure and geometric features of the Tibetan writing.We first let the system "learn" the geometric features of Tibetan consonant and vowel symbols,then split the word on the base of structures of print Tibetan.The results of the segmentation are used to do the next identification.At last,we convert them to Latin which makes people can "read" Tibetan.Through the experiment,we can see the accuracy of this approach can be over 89%.更多还原