摘要
在手写汉字识别的研究中，鲜有研究者提出建立手写汉字的数学模型，本文在这方面作了一些探讨。建模的目的通常有两个：一是手写汉字的表示或描述，二是手写汉字的识别。本文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。实验表明，该模型用于识别。
In the research of handwriting Chinese character recognition, no researcher raised a mathematical model of handwriting Chinese characters which is discussed in this paper. Generally, there are two purposes to raise a model: First one is for description, second one is for recognition. This paper is focus on small set character recognition. Based on skeleton graphics, handwriting Chinese character is defined as a set of isolated branch, isolated loop and component. Relation among these three elements is also defined. Then, mathematical model is raised. Experiment shows this kind of model is effective when used for recognition.
出处
《中文信息学报》
CSCD
北大核心
1999年第6期34-40,共7页
Journal of Chinese Information Processing
关键词
OCR
字符识别
手写汉字识别
建模
OCR Character recognition Handwriting Chinese character recognition Modeling
