【摘要】 提出了一种获得汉字笔画种类及其拓扑结构的新方法。通过分析TrueType字库存储的汉字字形信息,构造笔画图形并提取出笔画特征,采用统计分类和边界关键点定位相结合的方法,利用综合分类法实现了对二级6763个汉字的笔画种类和拓扑结构的基于XML的语义知识表达,可结合前端输入程序进行快速搜索和定位。为字体变形、字体自动生成、汉字知识挖掘等提供了必需的基本信息。实验表明这种方法能够准确、有效地识别汉字笔画及其拓扑结构。更多还原

【Abstract】 Propose an approach to classifying strokes and getting topology of Chinese characters.Shapes of strokes are constructed and characteristics of which are extracted through analysis of the TrueType font glyphs containing characters information.With compositive classified algorithmic,XML-based semantic knowledge representation for stroke types and topology of 6763 characters is implemented by integrating four kinds of statistical approaches and layout of core points.This method supplies fast searching and locating by using foreground application,and supplies prerequisite basic information for glyph transmutation,automatic typeface creation and knowledge mining of Chinese characters.The experiments show the accuracy and efficiency in recognition and topology of strokes.更多还原