小类别数手写汉字识别

SMALL SET HANDWRITTEN CHINESE CHARACTER RECOGNITION

针对小类别数手写汉字 ,在骨架图形的基础上 ,把手写汉字看作孤枝、孤环和部件的集合 ,并定义三者之间的方位关系 ,从而建立手写汉字的数学模型 .基于这种模型 ,进一步探讨一种新的识别方法以及新方法所使用的知识库的构造方法 .实验表明 。

Based on skeleton graphics, a handwritten Chinese character is defined as a set of isolated branch, isolated loop and component. Relation among these three elements is also defined, and then a mathematical model is raised. Based on this kind of model, a new recognition method, as well as the construction of knowledge base used for the method are. The experiment shows that the model and method is effective in small set Chinese character recognition.